Is Uniswap LP actually better than just HODL-ING?

Over Q3 2020 all of the decentralized finance (‘DEFI’) crypto-exchanges started growing rapidly, essentially as a result of Uniswap launching a more resilient and deeply-audited version of their exchange platform which had been working since 2018. This was combined with the ‘alchemy’ of monetary easing, fiscal stimuli and everyone being in lockdown, spurring the explosive results in the rest of the year.

The resultant influx of funds led to huge liquidity pools being created and many saw the opportunity to earn ‘fees’ by way of the protocol’s ‘non-negotiable’ automatic 0.3% commission, effectively allowing retail investors (not just the big guys) to ‘put their money to work’. With Uniswap pushing $4Bn in LP funds in early 2021 this is starting to get serious.

However, unlike a traditional ‘order-book’ method of providing liquidity in an asset, Uniswap and other DEFI exchanges use ‘automated market making’ (AMM). Uniswap’s AMM uses the ‘constant product’ model which relies on ‘arbitrage’ from external exchanges to close out the imbalances in the liquidity pools which are made up from two tokens.

Things you might have heard about LP and Uniswap

Its full of Scam tokens!

Well, unfortunately this is true. The ‘frictionless’ permission-less and almost cost-free ability for anyone to launch a token on the platform has prompted Uniswap to force acknowledgement of this by a warning disclaimer requiring ‘I understand’ before anyone puts funds in to LP for any pair. Many scammers are still creating ‘fake’ ERC20 tokens which otherwise look identical to well-established ones, luring investors in, before dumping these tokens via (perhaps pre-loaded?) fast transactions and cleaning out the ‘genuine’ half of the liquidity pair.

You always lose money when the prices of the tokens move

Again, this is essentially true. Since the AMM relies on ‘arbitrage’ to close out the imbalances in the pool, you as the liquidity provider are taking the ‘wrong side of every trade’. (why would anyone EVER want to do that!?). The good news is that on every trade done in the pool, you are earning your share of the “LP fees” (0.3% of trade value) in new “pool tokens” which are then in turn being used to provide liquidity again the pool. This adds up over time and “compounds” meaning that eventually you should be making money versus just ‘HODL-ing. Well, its not quite a ‘foregone conclusion’ as we will see below and you can check for any time period and pair using the free Novum Insights “defi-calculator” . Depending on the pairs and the volatility you can very easily lose money, as well as make money, of course :-). We look at this in more depth below. Worth remembering that these are called ‘impermanent’ losses as they may reverse and turn into gains over time as you accumulate fees and the prices move back. Or they may not. Also consider that when you are ‘winning’ these are impermanent gains which (even in 50% stable pairs) can vanish over night (the joys of a real 24/7/365 global market!) as all of your money is locked in ‘pool tokens’ and not actually in either of these actual tokens or indeed in your hands.

One of my pair is ‘stable’ so I am 50% protected from price drops

So there is some truth in this but it is not as simple as ‘you are covered. 50%’ We need to consider what happens when you are in a ‘stable/volatile’ pair and the prices move a little and especially when they move a lot. Ultimately if the value of the volatile side of the pair goes ‘to zero’ in market value, then the market will use your LP to trade all of your stable tokens for the ‘sh*t coin’ leaving the pool looking like the one here. We model this below to demonstrate the effects of volatility on this type of pair.

Gas costs make this unprofitable and when you need to get out — you cant afford it!

Gas did reach 600 gwei at one point over the summer and some token contracts use more gas than others so even just ‘getting out’ of a relatively mainstream pair like WBT/ETH would have cost over US$160. The prices are demand-driven and guess what? When everyone is trying to ‘use the exit’ all at once, the prices spike. You should do your homework and check whether can afford to ‘provide liquidity’ calculating all the trades you need to make and looking at a ‘gas site’ — like this excellent one here. A quick summary would be that 1.) if you are using big amounts of cash, Gas is a rounding error (lucky you!) and 2.) ETH 2.0 will fix this (we hope!).

Objective & Caveats

We set out to model the outcomes from some fairly standard investment objectives, summed up as;

Objectives

I want to get in to ‘mainstream’ Crypto — so I want to hold BTC and ETH (rather than lots of other things I have never heard of.)
I love the idea of my money making more money in fees while I sleep! — So ‘LP-me!’ and let the fees roll-in!
I am worried about it ‘all going to zero’ — so i am thinking one ‘stable-coin’ in my pair helps protect me from that?

Model

The model looks at the behaviour of token pairs (yes, not tokens) in liquidity pools under various ‘stress tests’ in both a negative and positive direction. If you are in a pool and one or both of the coins you are providing liquidity for ‘moons’ or ‘reks’ what impact does that have on your fees? portfolio value? If one half of the pool is stable, are you really protected or just handicapped?

We imagined that a new investor might want to hold both BTC and ETH and might also want some ‘hedging’ as an option with the use of one or more stable coins. We looked at ‘edge cases’ where one or both of the tokens went to zero or 3x and also looked at where one coin rose in value and the other dropped, even to zero.

We took an imaginary sum of $1000 and ignored all fees and split this into two pools one for BTC and one for ETH with stable coins (so thats $500 in each pool and $250 in each coin) or else the ‘moon-me’ all in option of $500 in BTC and $500 in ETH in one pool.

Another reason for performing this analysis and not some other stuff is that the two largest liquidity pools on Uniswap are ETH/stable and BTC/ETH

Caveats

We have ignored purchase, conversion/swap, permission, liquidity-add/remove costs/fees/gas — these may be significant(!) — do your own calculations on your amounts.
We have assumed that stable coins are 100% stable. This is clearly wrong as they fluctuate around the $1 (a little at least). As an example DAI has been up to $1.04 and down to $0.96 in the last 180 days, although this was rare and only lasted very briefly. So to provide meaningful analysis, we have assumed all of the stable tokens in our example stay at US$1. No matter what. They have so far but regulatory intervention or platform failure could change this (we hope not).
This again is a model and does not include ‘real world data’, from many more time/data points or include lots of other crazy coins and pairs but this is available from Novum Insights on request and might surprise you.
We assumed an annual ‘APR’ of 22% in terms of ‘LP Fees’ which may be massively over or under-cooked but there has to be an assumption of fees or else why are you doing this? Do your own research and remember of course that ‘past performance is not necessarily any indication of what may happen in the future’. In addition, the earning of fees and real world conditions of swaps, adds, removes and the dramatic changes in these which can impact the pool and pricing behaviour has also been ignored.
Assumed that ‘arbitrage works’ and in the case of major tokens it seems to work. Remember that in other tokens (especially the newest ones!) they may not be listed on other exchanges or have enough volume of transactions for the ‘balancing’ to work very effectively.
WBTC and BTC are exactly the same thing. They are definitely not and you should do your own research on this if it concerns you. More here.
All of this makes the assumption that nothing terminal happens to the Uniswap code either with the release of a new version or an upgrade to ETH2.0.
Most of the daily ‘action’ will be in a tight area of the chart around +/- 50% but over time and in exceptional circumstances ( i.e. any day of the year in crypto :-) ) we might reach the different extremities of the chart.

Results of differing analyses

1. Just HODL ETH and BTC

This is perhaps the easiest of all strategies to understand and does not involve Uniswap at all but we need at as a control before dipping our toes in the diving pool to compare.

There are three axes to follow; X is the relative value of BTC, Y is the relative value of ETH and Z is the value of your $1000 invested (net of all gas fees!). There is no real starting point but we should look at 100(%),100(%),$1000 (interactive version available from Novum Insights) as the point and then look at what happens as the token values move up or down relative to each other.

‘Just HODL’ BTC and ETH 50/50

Nothing to see here really. You get the gains and losses from movements in each of the tokens, nothing more, nothing less. So that would be $3000 if each of the $500 worth of tokens went ‘3x’ and $0 if they went to zero % of their original value.

2. LP both BTC and ETH, each paired with a stable coin

So this is where you hold $500 in BTC/USD(lets call it that) and $500 in ETH/USD. Which means 50% of your total LP is in stable tokens and the rest is in BTC and ETH. You get the gains (but not 100%) but likewise you get the losses (also not 100%) and if there is ‘divergence’ between BTC and ETH you are still in the game on both.

2 x 50% stable Pools of ETH and BTC

So in this scenario you are protected from all but the steepest drops and if one of BTC/ETH dies while the other moons, you still collect the moon earnings.

So the left and right ‘wings’ show this example for either of the coins and the peak at over $2000 where both triple in value. Pause there for a second. Both of my tokens tripled in value and my portfolio only doubled? Yes, thats right. Your ‘stable handicap’ was holding your gains back. Remember it is also (most of the time) giving you less severe losses. Likewise when both tokens go to zero in value, so does your portfolio (almost) as you get cleaned out as per the ‘50% question’ above.

It can be seen that this strategy effectively ‘hedges’ you if imperfectly(sic) from the losses as well as the gains and might be the right level of risk/reward for you. How does it compare to just HODLING?

3. HODLING vs two 50% stable pairs

In this instance, we are comparing scenarios one and two above which results in two ‘canopies’ one being the flat example in (1) and the other being the ‘winged’ canopy in (2).

HODL vs 2 pools of 50% stable

This is a bit harder to see without interactivity, but basically the ‘wings’ on (2) which includes 25% stable in each of the two pairs dip down on the edges of each range in the chart. Translated, this means that by LP-ing with 50% stable you are beating HODL-ing by earning and compounding fees, except where the movements are extreme when you lose out. You can try a few date ranges and see for yourself on this free calculator.

Worth noting that in the event of both BTC and ETH crashing, you are protected for a while by the stable effect, but then your portfolio drops rapidly ‘in the end’ as described above. You are dropping less quickly than just HODLing we should mention.

4. BTC/ETH in one AMM pair starting at $1000

This is the ‘moon-me’ strategy where you want to hold both BTC/ETH tokens but also get paid in fees for just holding the tokens (sounds great!). That surely must be better than just keeping them like you have been, right? Well, we shall find out.

BTC/ETH in one AMM Pool

Wow! A lot more vicious than I thought at least. I expected it to be the ‘best of both worlds’ or something like that? The wings are a lot steeper on the downside than in the ‘stable’ scenario which I guess makes sense since if anything too drastic happens to one of the coins you are holding, the AMM and arbitrage mechanics will ‘level it out’ delivering to you as the ‘bag-holder-in-chief’ all of the tokens which have died in exchange for your balance of the still-valuable ones. So how does this look versus just HODLing BTC/ETH? Let's see.

5. BTC/ETH in an AMM pair versus HODLing

This is scenario (1) versus scenario (4) where we compare AMM to just holding with two volatile (non-stable) tokens.

BTC/ETH LP vs HODL

Interesting. As long as both tokens are rising the AMM pool does better than hodl-ing but sudden dips in either or both very quickly penalise you as the market move is closed by arbitrage. This shows that in extreme cases ‘you get stuffed’. It is a general rule for automated market making, meaning at some point (unless you want to lose a lot) you would have to ‘get out’ which would cost ‘gas’, probably just when it costs the most. An interesting point on this is that unlike in traditional markets no-one can ‘refuse to fill your order’. It all just becomes about paying enough gas to have it mined (see gas above).

The other part which is predictable is that if both tokens ‘moon’ then you get ‘moon + fees’ which of course is more than ‘moon’ as long as there are fees and you leave it in long enough. It is then a perfect ‘bull’ liquidity pool meaning that as long as ‘price go up’ you will be ‘in the gravy’.

There is one final analysis to look at. How does the ‘bull’ strategy compare to the ‘50% stable’ in two pools.

6. Dual 50% (2 pools) v all in 1 BTC/ETH pool

This looks at the ‘dual-stabling’ of two pools as per scenario (2) versus going all in with the full starting value in one BTC/ETH pool. As an important point here, the ‘Dual 50%’ needs to be in two completely separate pools or else it just does not work and you are exposed as per one BTC/ETH pool.

2 x 50% Stable Pools BTC & ETH vs 1 x BTC/ETH Pool

And there we have it. Perhaps predictably, the 50% stable-ing and the splitting in to two pools protects against all but the most severe price drops. In the case of one token ‘dying’ then you still have the other pair. By contrast, when the ‘moon’ happens for both tokens you make a lot less. Of course, if both tokens go to zero, you are still cleaned out in the end as the tokens go to zero and your ‘stables’ get arbitraged away into the ‘ether’ (sic).

Conclusion

In summary, it is possible to ‘provide liquidity’ to the market and make more money than just ‘HODL-ing’ in the long run but not under all circumstances. It is also possible to ‘configure’ your LP portfolio according to your risk-appetite and opinion (guess!) on what the market will do next. Stable coins in your pair do provide limited protection from sharp price corrections but also limit your earnings on sharp price increases. More information, analysis and interactive charts as well as the defi-calculator available at www.novuminsights.com.

The information provided in this article by Novum Insights is for informational purposes only, we make no warranties of any kind, express or implied, about the completeness, accuracy, reliability, suitability or availability with respect to the article or the information, products, services, or related graphics contained on the website for any purpose. Any reliance you place on such information is therefore strictly at your own risk. None of the information provided is intended nor should be relied upon for the purposes of investment.

The ‘defi-calculator’, including the ‘top ten’ constitute entirely estimated historical returns which provide no guarantee, promise or calculation of potential future returns or losses. In addition, all of the assets displayed and shown on the website are highly volatile and risky.

Credit: Written by David Henderson for Novum Insights. Thanks for modelling and technical analysis done by James and Thomas including interactive 3d-graphing the scenarios. I could not have done it without you!