The practical pitfalls of trying to buy all the top cryptocurrencies
On the many, many hidden difficulties facing a seemingly-simple investing strategy — and the solution we cobbled together
In the last few posts, we’ve found that just buying and holding a broad, diversified basket of cryptocurrencies should (in theory, at least) help you get strong returns while minimizing your exposure to the idiosyncratic risk of any one coin. You don’t have to try to time the market (which is hard, and tends to go badly) and you don’t have to try to guess who’s going to win (which is especially hard in a fast-changing world like crypto).
This theory makes a lot of sense to me, since the same logic applies to the stock market. But if there’s one thing I’ve learned about crypto, it’s that our shiny theories often run headlong into the messy, confusing, rough-around-the-edges reality.
So, my question became: while it sounds simple to just buy the top 100 cryptocurrencies, how easy is that to do in practice?
To answer that question, I pretended I actually had a pot of money to invest and sketched out every single step I’d need to take. Here’s the summary of this journey — and boy, was it far messier and harder than I thought it would (or should) be.
A quick recap of our investing strategy
Before we go on, I wanted to recap the strategy we’re going after, in case you haven’t read the previous pieces on this site yet.
So far, we’ve found two solid, index-based crypto investing strategies. First, we constructed a market-cap-weighted index of the top coins (which we called the Moon & Rug index), which had pretty solid returns: over 150% last year. Then we constructed a sort of equal-weighted index,1 where we “spent” an equal amount on all of the top coins. That one did better than the capitalization-weighted index, with the best one (buying all of the top 100 coins) returning over 540%.
So for now, we’re going to move forward as if we were trying to actually buy all the top 100 coins in equal amounts. But this methodology would work pretty much the same if you were going to buy more of the bigger coins; either way, you’re still trying to buy a wide range of cryptocurrencies.
And also note that we’re excluding all stablecoins (like Tether) and derivatives (like Wrapped Bitcoin or the wild leveraged version of Ethereum). The reason here is that stablecoins are pegged to fiat currencies like dollars, so you’re never going to make any money from holding them, and that including derivative coins would cause us to effectively double-count certain cryptocurrencies. So whenever I say “the top 100 coins” or anything like that, I mean “the top 100 non-stable, non-derivative cryptocurrencies.”
Not so Coin-basic after all
Anyway, the first thing I did was log onto Coinbase, which I knew sells over a hundred cryptocurrencies. Should be easy, right? Just plug in some dollars and buy ‘em all.
When I looked at the list of coins for sale, I noticed a lot of the usual suspects — Bitcoin, Ethereum, Dogecoin — but also some strange coins I’d never heard of. And, given that I’d just spent a while cataloging the list of the top 100 coins, I got the sneaking suspicion that a lot of the top ones were completely missing.
Turns out I was right: Coinbase has most of the top 20 or so coins, but beyond that it’s got some serious holes: it covers less than half of the top 100 coins by market cap! As we saw last time, it’s generally better to hold more coins (including those outside the top bracket) in your portfolio, so Coinbase was a disappointment on this front.
Very well, you might say. It’s still OK to buy just the top 10 or 20, right? Well, for one, you still won’t get all of the top 10 or 20, but even if you’re satisfied with a limited set of coins, Coinbase comes with another big problem: its fees.
Normal Coinbase (the simple UI most people use) has remarkably high fees. The lowest possible fee is 1.5%, which is only possible if you’re buying $200 or more in one go. If you’re only buying $100, the fee is 3%; if $50, then 4%; if $25, then 6%; if $10, then 10%. And this isn’t even mentioning the “invisible” fees of bid/ask spreads, which might tack another percent or two on. In a world where you can buy stocks for practically nothing (plus spreads), this is not a great situation.
Now, there is Coinbase Pro, the advanced UI with lower fees. But it still has some of the highest fees in the business, at 0.5% for anyone who trades less than $10,000. (And you have to deal with the spread.)
So, while Coinbase is a decent one-stop shop, it alone can’t let us implement our buy-’em-all strategy — and, even for the coins it does support, it might get outclassed by other exchanges with lower fees.
Other exchanges to the rescue?
Naturally, the next step was to figure out if any of the other crypto exchanges covered Coinbase’s holes. So I checked out two other popular crypto exchanges, Kraken and Gemini, as well as this “crypto bank” I’ve been investigating called Nexo. (You can buy and sell coins on Nexo, though it’s not a true exchange since there’s no open marketplace where you can trade with other traders. Instead, you just buy and sell from Nexo’s retail counter. But still, this works for our purposes.)
The bad news was that each of these platforms had significant holes. This is a snippet of the table I constructed to visualize the platforms’ coverage of the top coins:
So, every exchange has some serious holes: Kraken has Coinbase beat when it comes to Terra Luna ($LUNA), but Coinbase beats Kraken on the Crypto.com Coin ($CRO). The good news, though, is that when layered together the four platforms do a reasonably good job of covering the top coins,2 like how if you stack up slices of Swiss cheese, each slice covers the others’ holes.
So, a four-exchange strategy is certainly better than going all-in on one, though it’s still far from flawless. Even with all 4 exchanges, you still can only buy 56 of the top 100 coins. Imagine pulling up Fidelity, Vanguard, Robinhood, and Charles Schwab and not being able to buy 44 of the top 100 stocks from any of them. It would be madness!
Not quite what it says on the tin
Our seemingly-simple strategy of just buying the top 100 coins has run into some serious trouble so far. The best we can do is to sign up for four exchanges, and even then we can barely snag half of the top 100 coins. Much easier said than done, eh?
But we’re trying to make this exercise realistic, so we’ll soldier on. At the start of this experiment I only had accounts on Coinbase and Nexo, so I had to go sign up for Kraken and Gemini. As usual, I had to go through the process of uploading pictures of my ID cards and other biodata — the famous know-your-customer (KYC) ritual.
That wasn’t a big deal, but I quickly realized that each of these new exchanges were far from easy places to do business.
Kraken
Kraken has a nice UI, and you can buy coins instantly once you link your bank account, but this “Instant Buy” charges nosebleed fees of about 2%: 1.5% for the Instant Buy feature itself, plus 0.5% for using the automated-clearing-house (ACH) bank connection method. Yikes.
The lower-fee alternative is to fund your account yourself and then use their “advanced” interface, Kraken Pro, to buy and sell coins on the open market. The fees here are better, at a max of 0.26% (plus spread, but that’s always around, so I’ll stop mentioning it).
The problem here is that funding your account is a pain: Kraken doesn’t support free ACH deposits, so you have to wire money in — which costs $25 or $30 at most American banks. If you’re only trying to buy a small amount of crypto, this represents a massive fee. (Another alternative is to buy USDC on Coinbase, which is fortunately free, then to transfer it over to Kraken and sell it for dollars. But then you incur the ~$10 fees to send USDC, plus the small fees of selling USDC on Kraken. I think it ends up being slightly cheaper than wiring, but not by much.3)
So Kraken is fine, and covers a few coins Coinbase doesn’t, but the fee situation is a pain. Much less seamless than advertised!
Nexo
Nexo has no trading fees, which is nice, though I hear the spread is pretty bad. But, like before, funding your account is a pain: they don’t accept free ACH deposits, so you either have to spend $25-$30 wiring money in or maybe $10 sending USDC over. It’s not awful if you’re trading large amounts, but for a casual investor it’s going to be a lot.
Gemini
Gemini’s advanced ActiveTrader product, which is a lot like Coinbase or Kraken Pro, has modest fees, capping out at 0.35%. And it supports free ACH deposits, though for me that’s just $1000 a day, which is probably fine for small traders but I’m sure would get annoying for anyone trying to do this at scale.
Anyway, the takeaway from all this is that all four of these platforms have a lot more rough edges and “gotchas” than you might think. It’s not for the faint of heart. And you still can’t cover all of the top 100 coins!
Another interest-ing twist
An additional complication is that some platforms will let you earn yield on some of your coins if you leave them there. This post is already getting a bit long, so I’ll leave the deep-dive for later, but suffice it to say that Coinbase and Kraken will stake your Proof-of-Stake coins for you and give you a few percent of “dividends” a year (paid out in the original currency). For instance, Coinbase gives you 4% on any Algorand ($ALGO) you hold with them, while Kraken gives you 4.8%. It’s free money, and you earn it automatically, so it’s not a bad idea to take it.
Meanwhile, Nexo and Gemini act as banks and let you deposit your coins with them; they’ll lend out you coins and give you a bit of interest on them. The advantage here is that they can lend out just about any coin, so you can earn interest on a wide range of coins, while only a handful of coins offer staking.
The problem with lending — and we’ll examine this more later — is that things could go wrong. The platforms don’t give many details, but generally it sounds like your loaned-out funds aren’t insured (Gemini says your funds are held with “trusted partners” but also adds that there’s no guarantee backing your money). So if there’s a “bank run” of sorts, you might be in trouble. Meanwhile, your staked funds will be fine as long as Coinbase or Kraken stay in business and the blockchains stay secure — and if either of those failures happen, you’ll have bigger problems than just losing a few staked coins.
The long and short of it is that loans add some risk for not-so-amazing returns (crypto, if it’s successful, should grow so much that the 2-4% interest will be negligible). So we should generally avoid lending, though staking is probably OK.
Finally, let’s “buy” some coins
At this point, we have the machinery to compute which coins are in the top 100 (or beyond, actually; we have the data to go to 200 or 500), and we have accounts on 4 platforms that’ll us capture a good chunk of them.
The last theoretical step before we buy is figuring out which platform we want to buy each coin on. Given what we know about staking, fee structures, and the relative difficulties of transferring money into each of the platforms, we can say that:
Don’t use the lending features, even if they’re available. This means that there’s no real reason to use Nexo, given that it’s easier to get money into Gemini (I think Gemini’s modest fees cancel out with Nexo’s apparently-higher spreads). Note that there is one coin that Nexo has but none of the others do: Nexo’s native coin, $NEXO (ranked #84). I don’t think it’s worth wiring money in to get just that one coin.
I’m not sure about how expensive it’d be to send USDC to Kraken and sell it off for USD (and I can’t experiment with this without spending money). So we’ll have to pick between wiring or Instant Buying. Our fee will thus be either $30 (for wiring) plus 0.26% (for Kraken Pro fees), or a flat 2% (for Instant fees), whichever is lower.4
If staking rewards are available,5 buy the coin on Kraken or Coinbase, whatever has higher returns. (In practice, Kraken has better returns on everything.6)
If no staking rewards are available, get the coin on Gemini if it’s available there, given that Gemini has lower fees than Coinbase. If the coin isn’t available on Gemini, get it on Coinbase. If it isn’t available on Coinbase, get it on Kraken. (We’ll use the low-fee Gemini Pro and Coinbase Pro interfaces, and whichever Kraken interface matches the fee structure we end up going with.)
If we use this (admittedly complicated) logic, we come up with this breakdown:
I expected Gemini to make up the bulk of our volume, but I was surprised at how little we’d be using Coinbase. I suppose the platform’s mediocre rewards and relatively high fees are to blame.
Or, if we really wanted to buy 100 coins…
The problem with restricting our view to the top 100 coins is that we could only end up buying 45 of them. What if we really wanted our portfolio to include 100 coins? Well, we’d have to look beyond the top 100 — per my math, we’d have to look at the top 221 coins. Of those, exactly 100 would be purchasable on the 3 exchanges we’ve chosen.
Interestingly, Coinbase becomes more useful here, perhaps because it offers a lot of those lower-tier coins that nobody else does.
An advantage of this latter strategy is that our portfolio will be broader. The problem is that some of these tiny coins might be so thinly-traded that there wouldn’t be enough liquidity to sell off the coins when needed. Plus, the low volumes might mean a higher spread, though when I checked a few random small coins the spreads were less than 1% (in one case, just 0.1%).
So, if we’re dealing in small amounts of money, I think it’d be fine to expand to the top 221 coins so that we can actually put 100 coins in our portfolio. If we were doing this index fund at scale, though, we’d want to limit ourselves to the slightly-bigger coins. (Imagine having $100M of assets under management and having to buy $1 million of a coin whose total daily volume is $2.6 million. A bit iffy.)
Zooming out and looking ahead
From all of this research, I think we’ve found that it’s certainly possible to buy a small amount of all of the top cryptocurrencies, though it’s much more painful than it might seem at first glance. You can imagine how powerful it’d be if someone abstracted away all this work and offered a single index-fund product.
Going forward, there are a few more threads I’d like to pull on:
How and when you’d rebalance the index funds (or if that would rack up so many fees as to be a bad idea)
Whether staking and lending are actually good ideas
How much of a problem liquidity would be in practice
How you could combine index investing with dollar-cost averaging
Whether other weighting strategies, such as Penrose’s square-root method, would be useful alternatives to the methods we’ve evaluated so far
(Let me know if you want to hear about any of these in particular!)
But hopefully this provides a strong foundation for future work in the crypto index-fund space. Thanks for reading.
Thanks to David Kahn for helping with this research, pointing me to new crypto trading resources, and helping me clarify my investing philosophy. And thanks to my friends at the FLUX Collective for pushing me to sharpen my analysis and catch things like the liquidity problems I mentioned above.
The difference between this “buy an equal amount of the top coins and sit on it” strategy and a true equal-weighted index fund is that we aren’t rebalancing the portfolio. I promise we’ll get to that at some point.
One top coin that’s conspicuously absent from all the top platforms is the Binance coin ($BNB), but that’s because the Binance company and platform are restricted or banned in various ways in the US.
An even more clever alternative to wiring money is to send Bitcoin! Most people just HODL their Bitcoin or trade it off-chain these days (ironic), so Bitcoin’s network is uncongested and its fees are surprisingly low. In one test, Coinbase quoted me a price of just $0.13 to send Bitcoin! The problem is that you’ll have to eat transaction fees while buying and selling Bitcoin, so this strategy only really makes sense if you already have a large stash of Bitcoin you’re happy to part with.
The breakeven point is $1724. Below that, the total Instant fees are lower; above that, the total wiring fees are lower.
Note that several platforms let you earn staking rewards on Ethereum. This sounds juicy, but the catch is that they convert your funds to the as-yet-unreleased Ethereum 2, and you won’t be able to access them again until (or if) ETH2 launches. I have no idea when that’ll be — Ethereum 2 has been six months away for the past six years, as the joke goes — so I wouldn’t feel comfortable locking up my Ethereum coins indefinitely for a few percentage points of yield.
There’s a bit of a judgment call here, though. For some coins, the difference between Coinbase’s reward and Kraken’s reward is very slight (e.g. Algorand gets 4.0% on Coinbase and 4.8% on Kraken). If you have to buy Kraken through the Instant strategy, then Kraken’s higher fees might outweigh the marginal yield improvement. In other words, in some cases you might want to go for Coinbase anyway, even though it has lower rewards. I think the difference would be only a few dollars here or there, though, so it would be a wash in the long run.