LightStream Research was founded in 2017 with the goal of melding traditional research with alternative data and artificial intelligence to provide deeper insights into companies, industries and economic conditions. HFC’s Stefan Nilsson checked in with LightStream Research founder Mio Kato in Tokyo to find out more.
You launched LightStream Research in 2017. What led you to set up your own business?
I think it was mostly a desire to be able to focus my time on things which seemed useful and were simultaneously enjoyable to me. Equity analysis is simply something I enjoy doing and my education also heavily focused on mathematics, economics and econometrics, so the new big data and alternative data wave seemed like it was right up my alley. My time as part of company management at the firm Uzabase also made me realise that I like developing and training talented juniors. Lastly, I believe there is a significant advantage in being able to tap the lower wages and financial skills of labour on the Indian subcontinent but there is some rigidity of contract terms and high fixed costs for traditional KPOs. The confluence of these three factors made me believe that there was scope to develop a small team of talented juniors in Sri Lanka to emphasise data-driven research and flexibly provide investors with analytical help, both through standard research notes and bespoke research.
LightStream is part of a wider group of related companies that cooperate on different levels. What advantages does this set up give you over some of your competitors?
We partner with Nowcast under the umbrella of the Finatext group. Nowcast is Japan’s foremost alternative data provider and together with Sandalwood one of the few go-to companies for alternative data in Asia. We sit in the same office as Nowcast and collaborate with them extensively so we have access to a variety of credit card, point-of-sales and point card data. Naturally, this gives us some unique insights about various consumer product and ecommerce trends which are important to investors, but the longer-term picture is really about combining the skills of equity analysts and data scientists to convert very interesting data into accurate alert mechanisms to highlight investment ideas and also help build real conviction. To this end, we started an initiative to have all our equity analysts trained in Python programming, not just for the purpose of web scraping and data crunching but also to try and understand the way our Nowcast colleagues come at a problem. That deepening of collaboration will be a key determinant of generating the most actionable and high conviction insights in my opinion. We also publish our research on the Smartkarma platform. While we have no capital ties with them, we know the founders well and have a strong regard for the direction they are trying to push investment research in. It is early days for these online research portals, but there are a number of extremely experienced and capable analysts on the platform whose expertise we can tap. We are all about collaboration and trying to understand as many different perspectives on a potential investment as possible so we are very focused on trying to generate conviction on investment ideas by passing them through multiple filters – alternative data, traditional fundamental analysis, the perspective of industry experts outside of the investment world.
Your business and your research analysts are split between Tokyo and Sri Lanka. What was the thinking behind this set-up?
Mainly cost and the ability to provide an attractive opportunity to talent in Sri Lanka which is interested in cutting-edge analysis. Many KPOs focus on the cost advantage of analysts in Sri Lanka or India, but from my perspective what is more important is the ability to select the very cream of the crop when you don’t have to stress over budget. Conceptually, rather than trying to hire good analysts for a fraction of the cost, we want to hire the very best analysts in Sri Lanka for what is by Tokyo standards an extremely reasonable cost, but quite generous pay by Sri Lankan standards. We believe this better maximises the cost/benefit ratio. Also, I am half Japanese and half Sri Lankan and I grew up in Sri Lanka so I have an interest in trying to provide talented Sri Lankans with similar opportunities to what I was luckily afforded in my career. The cost advantage does, however, have important implications for the scope of what you are able to do. For example, with our Nowcast collaboration, developing entirely new methods of analysis and combining data analysis with the understanding of company business models and operating leverage is time-consuming. It would not be easy to make the necessary investments for this process if you were to hire only Japanese analysts, but by creating mixed teams it is much easier to obtain the manpower while still ensuring an understanding of the local dynamics.
You have a focus on Asian research rather than global. What made you decide on a regional focus?
Well, my own experience is in Japan where I have been analysing stocks since 2004. I also see a lot of positive developments in the Japanese market and among Japanese businesses so given the gradual migration of analyst talent over the last decade plus I do not consider this a bad time to be focusing on Japan. For Asia more broadly, we largely focus on three areas: 1) industrials, where China is often the marginal demand driver and where Japanese companies are particularly strong, 2) tech where Asia is often a key source of demand and very dominant in the supply chain, and 3) consumer products and retail, mainly focused on Japan where we most heavily utilise data from Nowcast. Since a lot of incremental global growth comes from Asia and we believe that Japan’s companies are not just healthier balance sheet-wise but also increasingly successful at expanding in Asia, and profitably no less, we believe that the key swing factors which drive performance will be frequently located in Asia.
Data and artificial intelligence are all the rage now. How are the data revolution and changes in how data is collected, analysed and used impacting your business?
It certainly feels like the future. On a personal basis, I think that the increasing abundance of data as people start to understand its true value could eventually lead to a greater emphasis on long-term success and strategy as alpha becomes harder to generate purely from data sets, therefore we are not by any means de-emphasising the importance of company strategy and qualitative analysis. With that said, in the medium term, we believe data analytics could become a very strong generator of alpha. In particular, we believe effective forms of data cleansing, which Nowcast is expert in, and generating high signal-to-noise ratios will be key. With that said, data quality is an issue throughout the industry and making effective use of a lot of the data coming through is not easy. At present data collection and tagging is not done well by all generators of data and improving this will be a process. Once that can be done it is a matter of integrating the much more granular and high-frequency data into financial models to parse the impact on company earnings. We are also quite fond of very traditional numerical analysis such as correlation analysis and doing specific factor analysis of company earnings. It is simple, but often effective and now we have even more data to play with. Manipulating and combining data in appropriate ways to generate investment insights also requires a certain amount of understanding of finance and equity analysis which many data scientists lack. Thus, we believe collaboration is necessary and easier said than done unless priorities are appropriately aligned. Given that the alternative data industry is just starting to come into its own, best practices have yet to be established. Hopefully, LightStream and Nowcast can combine our expertise to do that.
You have worked with analysis and portfolio management in the Tokyo offices of a couple of international alternative investment funds in the past. How does this hands-on experience help you in working with fund managers as clients now?
I think the biggest advantage is understanding what they want, what they value and how they think. I prefer to think of LightStream as an outsourced buy-side analyst shop rather than a sell-side research provider or even independent research provider. We want to generate ideas that generate good returns and we want to effectively identify risks to the trade and be very aware of the path risk for stock prices. I personally do not consider myself a particularly good or bad trader, but I consider it important to try and accurately frame the catalysts and risks of an investment case to make the life of someone who is a good trader as easy as possible. I consider the analyst side of the work to be my calling but portfolio construction and risk management is also a very interesting discipline and intellectual challenge. Being conversant in the latter I feel helps when articulating ideas to PMs and particularly so when you are talking to many PMs who could have very different trading strategies, risk tolerances and investment horizons. My own bent is to look for ideas which are structurally sound and likely to be attractive on the long side for 1-3 years, but within that time period desired sizing could change dramatically based on price performance and upcoming catalysts. On the short side, I find that effective time horizons tend to be shorter, perhaps 6-18 months. Of course, for alternative data driven ideas, trading periods could be much, much shorter depending on what the data is telling you.
When it comes to working with hedge funds and other alternative investment managers as clients, it is becoming more and more important to be able to offer a flexible and often customised service in order to win new clients and keep existing ones. How flexible is your offering?
I absolutely agree. Flexibility and adaptability are increasingly important given the changing landscape not just in the alternative investment management industry but in almost every industry globally. So much is changing that the ability to deploy resources flexibly and have access to a diverse range of skill sets is truly key. We feel that we work very well with funds that require ad-hoc bespoke research as we do not demand specific time period commitments. Many outsourcing companies lock clients in to year-long contracts and bill monthly so even if you don’t have specific ideas that you want work done on you are still paying for that time. We are structured to steadily conduct our own analysis on companies and publish this through platforms such as Smartkarma as our “day job” but have the flexibility to respond to urgent or short-term requests, especially regarding data analysis. Ultimately our client relationships operate on a trust basis and “selling” is definitely not our speciality. We thus focus on trying to be high touch but low maintenance. When our clients have a need we are there and happy to customise our offering according to the exact scope they are looking for. When they don’t, we are quite happy to be out of sight and out of mind until they next need us.