The Art Of Coin Selection: Building Smart Crypto Indices

Transcript

[Music] Good morning. Uh the Thursday morning talk is always an ambitious one. Uh so thank you very much for those who are in the room and I will try to make it energizing. Um my name is Subon. I'm the CEO of Kaiko and Kao is a market data company. I'll talk about that in a second. But today we are here to talk about the art of coin selection. What does that mean? Uh how do you build indices, crypto indices?

How do you build investment products that gives you exposure um to crypto as an investable asset class and how do you select those coins? So um this is what we're going to talk about today obviously with a very strong data angle. All right. So maybe just a few words about Kao first. Kao is a global company. We've been around for 10 years. Um and we basically work with businesses and financial institutions on uh their crypto data solutions and building data solutions specifically in the context of building financial products. Uh we're a global company. We're have offices in New York, London, Paris, Singapore, and Hong Kong. And we are regulated and audited and everything that you can see here. Uh what do we do? Market data. We

do analytics, risk management, quant solutions, we do monitoring solutions and we do KO indices and this is what we're going to be focusing on today. Now the first question is in order to say how do you build crypto indices the big question is which assets do you include and how do you go about the methodology about including assets in portfolios. Just to give you a sheer sense of the number of crypto assets out there. Um this is a chart that represents um crypto assets being listed on centralized exchanges over time uh since the vast what six years now. We have over 8,000 almost 9,000 crypto assets listed on centralized crypto

exchanges and that does not include the crazy amount and we're talking I think in hundreds of thousands of uh numbers of assets listed on decentralized exchanges. So we used the centralized exchange listing as a first proxy to say you know there is a first level of filtering and that is already almost 9,000 assets. So how do you filter that down? The methodology for Kao indices and for building um a selection of indices relies on four things. The first one is the exchange ranking. If you want to build an investable product and if you want to include crypto assets in your investment product, you need to be able to trade those assets on exchanges that are reputable. So the first part is the exchange ranking. It's a public

methodology. It's updated on a quarterly basis. It's entirely transparent and datadriven. And in order to have a crypto asset included in a Kao index, it needs to go through our exchange ranking. It needs to be listed on an exchange that is wellranked. Then there is the custody vetting. This is exactly the same ID behind it. It's to say if you want to invest in an asset you need to be able to custody that asset in a reputable custodian or to have a strong custody solution for those. Then obviously you have the liquidity assessment. Liquidity is extremely important uh because if you want to invest in an asset it needs to have sufficient liquidity that you can actually trade it and hedge it. And then finally taxonomy. Uh this is reference data. And this enables us to uh divide, sort, and filter all of these kind of

assets um that are out there, the 9,000 something assets into groups and subgroups based on their core nature and their usage. And that will enable you to build thematic indices and other kind of ways of filtering this um universe of crypto assets. So now um if we look at um the top 10 crypto assets what we see and it's quite interesting is that you know it's actually not that strongly correlated. Obviously the industry as a whole has its ups and downs and bull and bears but if you look at top 10 direct price correlation what you see is that bitcoin eth and soul you know top three are highly correlated between each other but uh if you go down into top 10 it's actually not that

correlated. What does that mean? That means that diversification is good in the context of building an investable portfolio of assets. Right? What you want is that if you have pure and perfect correlation, it doesn't make sense to have a basket index. Here what we're saying is building a basket and having um a selection of coins going beyond top three and also we will see later not too much further out of top 10 uh is actually very beneficial. It brings um it brings diversified exposure. Obviously, within the methodology, you're going to cap the weight that each asset has in a portfolio, but diversification is good.

And the other important part is the selection bias, right? If you invest in a top 10 crypto and you outsource the decision of your investment, you're going to remove selection bias from building your own portfolio. So, the big question here is everybody talks about Bitcoin. I come from the world of trfy and everybody all my former colleagues for banks they all say you know we invest in Bitcoin ETFs because it's safe and because we know Bitcoin is digital gold it's safe we know it we trust it it's now a you know an asset that is an acceptable institutional investable asset um but are are we ready to go beyond BTC um and clearly what we see now is that if you

want to get exposure to crypto um with a little bit more depth in how you invest in crypto. You now look at this as a growth sector and you need to not focus on single assets but build a diversified slightly deeper exposure to the asset class. Um in the context of portfolio constructions obviously basket provide a systematic exposure to that growth sector and this is uh something that will mitigate single asset risk and selection bias. Now, of course, from an institutional standpoint, it is now uh you know, index strategies are enabling participation in the sector innovation at large and the

market leaders beyond just Bitcoin and ETH. Uh and this is what we're trying to achieve with proposing a basket strategy here. So, the big question, how many crypto assets are investable? We clearly you know know that 9,000 is not the right number. What is the right number? How do we get to a number where we know there is sufficient liquidity, it is sufficiently institutionally investable and it is um safe to invest in as in you can custody it and you can trade it. So we go through this methodology uh where we start from the digital asset full universe of 8,000 something assets and we start removing all of the ineligible

coins and those are either coins that you know should not be part of a portfolio for reputational reasons but also all of the things that actually are not going to give you a crypto exposure either because it's a stable coin, right? You don't necessarily want to have stable coins in a crypto portfolio because by default it's a stable asset. you're better off holding your dollars. Uh, and we remove all of the kind of uh, direct stocks and everything. We want a crypto investable universe. So, the first part is removeable asset. We end up with a digital asset iligible universe that is about 1,500 assets. So, we've already skimmed down pretty much.

Then we start removing the um you know uh we start adding the filters of the exchange ranking, the liquidity vetting and we move down to this uh digital asset investable universe. Assets that we know are actually tradable and investable and we're already down to 150 assets. So we've significantly shrunk down the investable universe. And then we apply our own methodology and we end up with KT10 which is our Kai code top 10. And this is um what we deem is a you know safe institutional investable asset um uh universe and we have a top three, five and 10.

So uh one of the things that I've been fighting against over the past couple of years is the market cap trap. Very often when we speak to people uh that are not necessarily deep into the weeds of that, they look at market capitalization. They sort cryptocurrencies by market cap. And this has bugged me since day one because especially in crypto market cap doesn't mean much right because market cap you take the number of assets you multiply it by the latest traded price and boom you have a market cap. The reality is what we should look at is liquidity is um the natural organic market depth of those assets as they are traded on centralized exchanges. looking at order books, looking at like the market micro structure, looking at spreads, looking

at liquidity slippage, all these kind of liquidity metrics that are really a true reflection of an asset's organic liquidity and the industry's demand for that asset. Um, so we try to really, you know, avoid using or overweighting market cap in our methodologies and we overindex on liquidity metrics. Um so here what we're seeing is the evolution of top 10 uh assets by market cap and what we're showing is that it actually changes a lot. So if we only built methodologies on um a market cap metric in the how do we decide which assets are included in the methodology what we see here is that actually 28 different assets have been moving in and out of

top 10 by market cap. So this actually would be difficult to maintain if we were rebalancing on a whatever frequency basis whether it's monthly quarterly if we're rebalancing every time an asset move in and out of market cap of sorry of of top 10 by market cap this would actually increase costs and create slippage and actually create a bit of inconsistency in the methodologies. I'm not sure where I should angle this clicker to make it clicky. All right. Um here something else that we want to share is bringing you to the fact that we believe the right number of the investable universe should be somewhere around 10. Uh if you look at

93 crypto assets that we deem to be investable according to all of our kind of building blocks and liquidity criteria and the blocks I've mentioned before. What we're seeing is out of that only two are considered large caps. Obviously, uh those are Bitcoin, ETH and we get you know Solana in and out of that. Um when it comes to midcap, there are seven that we consider to be midcap and then everything else goes into what we call the small cap threshold bucket. So obviously two and seven are taken into our KT10 methodology and then um a mix of the smaller caps with different weights uh and into how they're included

in the methodology. But yeah, what we're seeing is that actually it's really hard to get further down um the you know number of 10 when you're building an index portfolio. And actually um you know there are a couple of indices uh there is a top 20 that is done by not Kaiko and there is a another one that is quite famous which is the Coinbase 50. Coinbase 50 currently includes something I think which is 20 to 25 assets. Uh the goal is to go to 50 but even they they're building and and kind of marketing a top 50 strategy. Um but they know that they cannot realistically include more than 25 assets today. So it's interesting to see it in the data and to see that if you're trying to build I think there's clear benefits in

diversification beyond Bitcoin as you're building financial financial investment products but going way beyond 10 doesn't make so much sense at this stage. Um so as I mentioned market cap very dangerous slippery slope and I think a a um you know a misleading um financial indicator for the health of uh and the investability of an asset. But we are focusing on liquidity and liquidity is not just traded volume. Liquidity is also um the availability of derivatives for those assets enabling hedging and making the actual assets uh more liquid and with alternative uh hedging solutions. So here we're looking at uh our our list of top 10 by liquidity

threshold by our own liquidity ranking. And what we see here is that all of these top 10 assets have uh access to both spot but also futures and options market in the US and in Europe. Uh and that is really interesting. I mean uh we have we work with one of our clients uh Gemini on the derivative side. They have listed our top five index per uh derivatives exchange in order to give exposure to a basket of five assets uh in a single investment product. So the only the availability of those derivatives uh markets for single assets is very important and we are uh working with derivative venues in listing per futures on basket strategies and this is

the basket that we're pushing which is a top 10 just check on time. Um, so now the big question, should memecoins be included? That is a question that you know depending on who you ask, you have a very different answer and I let you imagine the kind of answer you have depending on the kind of person you ask. So we've done a survey. We've uh um reached out to our 200 something clients and we asked, do you want us to include memecoins in our index methodologies?

And unsurprisingly, crypto people think that yes, we should include memecoins and trafi people think that uh you know, a majority of them think that no, we should not include memecoins. So, it's a very obvious result. Uh what I think is interesting is that in trfy I would have expected it would have been even more binary, even more like absolutely not. We don't want to see memecoins. And I think it's a good indication that institutions are starting to understand why memecoins matter uh and why they actually bring this kind of very different maybe you know almost digital unique to digital asset sense of community and belonging and network effect and there is something that is specific to memecoin. Uh obviously there is a long tail end of um maybe less

relevant memecoin but in our KT10 we include Dodge and uh Dodge is a kind of very powerful memecoin from a network effect standpoint and also from a price action standpoint. Actually we're going to see that next. So uh what we're showing here is um two charts. One of them is our KT10 with um memecoins. I mean I think Dodge is the memecoin the only one. and uh the same index without um memecoin. What we see here is that actually they don't always add outsized return and they don't always add crazy volatility either. Very high correlation obviously because of the waiting that never lets you know um

given memecoin take too much um too much weight in the given uh strategy. But what we're seeing here is that um KT10 without memecoin is actually very slightly from a price performance slightly above um KT10 which is interesting because very often when we talk to institutional client they think oh this is going to add crazy amounts of volatility and actually th those that are in favor think that it is a proxy for um the general crypto health and industry and activity and that thereby when uh the price goes up uh any index with memecoin would actually go up more.

What we're seeing here is that it's not necessarily the case. Um so now we you know I talked about the methodology of building indices. We've shown data around the the assets themselves. Now when you have a good methodology it's great but it's not good enough uh with uh those kind of things. Data is critical and it's very much garbage in garbage out or quality in quality out. So here um Kaiko indices is here on the right and I mentioned all of the building blocks that enables us to administer regulated benchmarks that is used for uh the purpose of building mono asset financial products and also Kao indices which is kind of basket indices.

Um but all these methodologies are great what you need to funnel through is good data. So for those who are uh you know familiar to the world of market data in cryptocurrency it's an extremely fragmented market uh centralized exchanges are uh numerous uh and uh you get market data which is like price data order book data and trade data from hundreds of different sources uh all over the world in real time. So you have to get all of this market data at the centralized exchange level and uh we are in a world where DeFi is I believe the future and so we are seeing very uh you know strong volume and growth on DeFi market activity. So obviously you cannot

ignore all of the DeFi market data liquidity pools and trade data happening on dexes. So there's a big a lot of work in you know being able to collect this data and and standardize the data and put it in a database so that it is ingestable on the output um regardless. The idea is that when you consume the data you don't have to know whether this trade um was from Coinbase or from unis swap. It's a trade and you need to be able to ingest that. So that is what Kao data does and Kao data powers all of the index methodologies.

And so what do you do with indices once you've selected your KT10 your top 10 indices? What do you do um with them? So couple of things you can build indices that are going to power ETPS and ETFs. Currently we work with uh most of the large crypto asset managers in Europe that are issuing crypto ETPs and listing them in Europe. Uh ETFs, we've seen a lot of monoasset ETFs in the US until until now. uh we are starting to see uh in the applications uh more uh basket strategies and basket ETFs and obviously structured products and all of the regulated derivatives. This is also important. I mentioned Gemini launching uh top five uh per earlier this year. We

are working with most of the regulated derivatives venues on listing um per on baskets and futures on baskets. This is also important because the second will have an impact on the first because the more um you know um derivatives and per you have on regulated exchanges on the same underlying that you are building ETFs and ETPs the more hedging capacities you have and this brings uh you know the entire industry up by providing liquidity and investability and more diversified exposures. All right so this is what we do. This is how we build indices. Uh the team is here. I'm going to be around. So if you want to follow up, I will obviously uh be available for questions. But thank

you very much. [Applause]