CME vs Polymarket Arbitrage
Disclaimer All content and information on this website (the "Site") is for informational and educational purposes only, does not constitute financial, investment, legal, or tax advice, and is not an offer, solicitation, or recommendation to buy or sell any security, futures contract, or financial instrument. No Investment Advice The author of this Site is not a financial advisor. The information presented is not intended to be and does not constitute personalized investment advice. Any discussion of securities or markets is for illustrative purposes only. You are solely responsible for your own investment decisions. You should not make any financial, investment, legal, or tax decision based on any of the information provided on this Site without first consulting a qualified and licensed professional and conducting your own research and due diligence. Risk of Loss All trading and investing involve a substantial risk of loss. The assets and strategies discussed on this Site may not be suitable for all investors. You may lose all or more than your initial investment. Past performance is not indicative of future results. No Warranties The information on this Site is provided on an "as is" basis. We make no warranties or representations, express or implied, as to the accuracy, completeness, or timeliness of the information. We are not liable for any errors, omissions, or any losses, injuries, or damages arising from your use of this information. Any forward-looking statements or market predictions are speculative by nature and are not guarantees. Disclosure The author may hold, buy, or sell positions in any assets, securities, or financial instruments discussed on this Site at any time and without notice. Personal Opinions All views and opinions expressed herein are the author's own and do not represent the opinions of any entity whatsoever with which the author has been, is now, or will be affiliated.
There's an interesting potential arbitrage opportunity on the markets that speculate on whether the Fed will cut interest rates in the next meeting or not (Prices at time of writting)
Polymarket is currently trading at ~70% implied probabilities while CME ZQZ5 imply a lower probability (way more liquid so that number fluctuates, but it is around 66-68%)
The obvious trade would be to go long the ZQZ5 contract while selling 70 cents of the polymarket bet, but the two instruments differ a lot in their pricing mechanisms. The bet is a binary yes or no, keep the money you received or pay back 1 dollar for every 70 cents you received, while the ZQZ5 contracts are priced according to the average of the Effective Federal Fund Rate (EFFR) of each month, which fluctuates
In order to understand this trade we need to model all the possible scenarios and prices at the date of the Fed decision (2025-12-10)
The price of the contract is given by the formula 100 - R with R being said average
The tricky (and risky) part of this is to try and predict the EFFR during the month of December
EFFR and ZQZ5
https://fred.stlouisfed.org/series/EFFR
Banks are forced to re-balance their cash reserves every night, and the EFFR is the average rate they pay each other.
The Fed steers this rate by setting two "boundaries":
The Floor: The ON RRP rate, which is the lowest rate the Fed pays to borrow money.
The Ceiling: The Discount Rate, which is the highest rate the Fed charges to lend money.
The EFFR stays stable just below the IORB (the high-interest rate banks earn on deposits) because of a simple, daily arbitrage trade
ZQZ5
https://www.cmegroup.com/markets/interest-rates/stirs/30-day-federal-fund.contractSpecs.html?videoId=6372864185112&utm_source=chatgpt.com
Each unit move (ie the price goes from 96 to 97) results in $4167 P/L (5M / (1/12) * 0.01)
This makes our P/L very sensitive to EFFR movements
In order to predict EFFR, we take a look at previous levels vs both the boundaries and the IORB
With data going back to 2020, the EFFR has never exceeded the mid point of the boundaries, with a recent median being around 30-40% of the gap (depending on timeframe window)
We also compare EFFR to the buffer (IORB) that the Fed offers (which is always lower than the upper bound), with a median being around 55 to 65% again depending on timeframe
The worrying thing is that the recent EFFR levels are at the highest they have been the past 5 years, sitting on 48% of the total spread between the boundaries
Modeling the payouts
Price of ZQZ5
Price = 100 - R
R = (10/31)*EFFRprior + (21/31)*EFFRpost
EFFR (adjusted for each period accordingly) = ( Upper Boundary - Lower Boundary ) * P ) + Lower Boundary
P = What % of the spread between the Boundaries EFFR will be at
Essentially P is the variable in which we speculate on by studying past levels
The lower P is, the higher our P/L gets, but its very sensitive on it
So, on day of decision if Cut:
Price would sit best case scenario (assuming P = 30%) at 96.3275, or 96.295 for P = 50%
And in case of no Cut, 96.175 or 96.125 respectively
With each point movement worth $ 4167, in the case of P = 50%, we earn $489 if the Cut happens and lose $218 if it does not
We now have to size our Polymarket Bet accordingly so that we create an arbitrage trade
The hedge ratio we will use that the price swing of the CME win/loss scenario, must reflect the one from Polymarket
With some simple math we end up calculating that for each CME contract, we need to be selling around 708 Polymarket bets
Lets see what ends up happening on the days of the decision
In both cases, we make $61
The amount we need to open this position depends on the margin requirements to trade these CME contracts and the amount we need in our Polymarket account to sell these bets
According to https://www.cmegroup.com/markets/interest-rates/stirs/30-day-federal-fund.margins.html?videoId=6372864185112&utm_source=chatgpt.com#sortField=volScanMaintenanceRate&sortAsc=true the CME contract requires a maintenance of $615 per conctract, but the final margin depends on our broker, so lets assume double that (needs further research on that part)
The polymarket bet will require to be able to pay the full amount in case we lose the bet, so lets assume that is a full dollar for each bet sold (also need to confirm this)
Thus the total position cost is assumed currently 615*2 + #of poly bets = 1938, and with a return of 61 in each case, we observe a potential ~3% arbitrage opportunity
If we lower P to our best case scenario, the return goes to a too good to be true 13%
In both scenarios we would exit the position on the day of the decision, giving us an IRR of ~30% (which also sounds way too good to be true)
Things to think about:
Liquidity, according to current Bid/Asks from Polymarket's order book, the biggest position we would be able to open on these levels would be around $130k
EFFR sensitivity and further understanding of each what can move it
Is pricing of the ZQZ5 contract modeled correctly - according to the research I've made it should be correct but some too good to be true results are worrying me
Finally let's also test our worst case scenario possible, which is that EFFR = IORB (in absolute worst case scenario that will be 3.7 (has never happened - and realistically cannot happen)
in that case our P/L looks like this
A realistic worst case scenario should be that EFFR goes to 55% of the spread (currently sits at 48%, has never jumped more), in which scenario our P/L looks like this
After 55% we start losing money on the trade
EDIT: Position required for the Polymarket Bet is actually 0.3 for each open position, not the full amount, making our required capital less and returns slighlty higher
Comments
Post a Comment