Will C2 ever support attribution of carry interest cashflows for fx systems?
Whenever you hold a ccy position, you earn or pay an interest rate differential. This can influence the PnL quite heavily, depending on the system.
This also leads to the second question, will any other pairs be added on C2 (EUR/TRY, USD/TRY, etc.)?
Any news on the carry interest question?
I could use some help on this: does anyone know of a machine-readable data source that I could use to help me calculate respective currency carry costs? Can you, Trendy, or anyone else, volunteer to design the necessary algorithms once the data source is located? I am happy to implement the feature, once a data source/algorithm has been selected.
I am not a forex trader, but is something that is going to vary widely, based on the broker? or instrument? if so, this might have dubious value.
in futures trading, some brokers let you purchase t-bills to earn some interest. some do not…on large accounts, the extra few percent might be significant. but it would be more trouble than it is worth to try and reflect in C2.
The way it works is that each broker sets their own interest rates. Also, the rules of when to credit/debit carry interest are typically broker-specific.
For example, Oanda calculates interest every second and credits/debits it either immediately after reducing or fully closing a position or, if the position is kept past 5 PM EST, they automatically credit/debit at 5 PM EST. This is regardless of whether it’s weekday, weekend or holiday. Their rates change in line with benchmark rates and they publish them here:
Their calculation policy is fully documented here:
FXCM calls the interest credit/debit operation “rollover”. Their policy is here:
Rates from FXCM show up on their platform in the “Roll S” and “Roll B” columns. Any position open or existing at 5 PM EST is subject to rollover. The credit or debit is posted to the account within an hour of 5 PM. The rollover doesn’t happen on Saturdays and Sundays and holidays but the amount to be debited/credited isn’t simply forgiven, it’s just applied the next day.
Carry interest is applied across all retail brokers as cash towards their account’s balance.
In case of FXCM, the “rollover” (interest) rates most likely should be present in the price feed you get from them, since the Dealing Rates window shows them among other quote details.
so, how is C2 supposed to account for each broker with one formula??
The reason why carry interest is significant lies in the same principle as the carry trade - interest rate differential. Borrowing one currency (to pay low interest), converting and investing in another currency (for simplicity: e.g. buying government bonds) allows the investor to not only profit/lose from the currency rate fluctuations but also from the difference in interest rates.
In the spot fx market, this can translate to not just significant cashflows but also to entire strategies.
Using bid/ask rates from Oanda, consider the following. When borrowing Yen (short JPY), one has to pay 0.8%. When depositing Australian dollar (long AUD), one receives 6.95%. Net result is that the person receives 8.15%. When divided by 365 to get an approximation of a daily interest, that’s 0.0223% per day, which in case of a 1M AUD/JPY long position translates to an inflow of 223 Australian dollars.
Given the miniscule returns that typical reliable FX strategies can achieve, an extra 8.15% can make a big difference. Obviously, this interest rate differential doesn’t come for free. The holder is still exposed to the currency rate fluctuations.
However, to credit/debit carry interest is a matter of fairness because it works both ways.
If we were to go short AUD/JPY, we must pay 7.325% on AUD and we receive 0.25% on the JPY, netting the interest to be a payment of 7.075%.
Understating or overstating PnL due to cost of carry would be a significant flaw in PnL reporting and no serious trader, fund, bank or broker neglects this.
Ultimately for investors looking to participate in C2 systems, carry cost can make a difference between a profitable or an unprofitable strategy.
I would suggest that C2’s policy follows the least common denominator practice. I also believe it needs to be instrument senstitive.
I.e. spot FX instruments should be all assigned to follow one policy.
As far as actual rates, every broker ends up ultimately marking up LIBOR (or other *BOR rates for different countries) or the corresponding central bank’s rate. A quick survey of the brokers will reveal, based on their rates, what those spreads typically are. For example, Oanda pays 4.95% on long and charges 5.15% on short for GBP. The BoE base rate is 5.0% and the 1-month GBP Libor today is 5.39%. I personally have a feeling that retail brokers set their rates based on central bank rates not based on any other money market rates because then they’d have to adjust them daily.
So it becomes a task of setting up:
1. instrument groups (assigning instruments to them, e.g. “spot fx” containing all spot fx pairs, etc.)
2. policies (defining calculation policy - time of day, daycounting convention for acrual, etc.)
3. rates tables (defining where rates are coming from per each country)
And then mapping each instrument group to a policy and a rates table.
This will be broker agnostic but would reflect conventions commonly applied by the brokers across different instruments.
Sorry, I’ve got some accidental numerical calculation errors in the previous examples but hopefully, my point gets across.
as I said, I am not a forex trader.
But when doing the major forex pairs, do you basically receive interest on one side and get charged on the other?
So for all intensive purposes, does it become a wash?
I am not seeing why this is significant.
No, it is the net differential between the two rates and can be highly significant. There is an excellent tutorial here:-
Basically for several years now most forex pairs effectively pay you to be long the dollar because of the differential in interest rates. And obviously it works the other way, so a forex system that held short dollar positions for long periods of time in the past would have actually given a portion of it’s profits back in interest payments, but C2 doesn’t reflect this. To be fair, this is the long term forex equivalent of the slippage and commission costs debate on scalping systems, any subscriber that was following a forex system that was often short the dollar and always held overnight would have easily underperformed the C2 system stats in their real life account because of the interest payable, it’s quite feasible to have lost money on a trade that C2 showed to be profitable, so it is definitely something I believe should be looked into. I would have thought the most obvious source is to try and use one or a combination of the brokers currently available for forex autotrading, those are afterall where the actual rollover / interest payments are calculated that people will be making or receiving in their real-life account.
Yes, in FX trading, whenever you have a long position, you are being long the base ccy of the pair and short the counter currency of the pair (or vice versa for shorts).
It’s significant because if you are strictly a spot fx trader and you don’t close positions by the end of day, you need to roll over your spot position to the next business day. It comes down to the intricacies of the “spot” instrument (which carries a T+2 or T+1 settlement, depending on currency pair) which are analogous with overnight borrowing.
The interest is not insignificant and affects the PnL. Propertly accounting for it also contributes to being much more realistic about the strategy’s success should you ever scale it up.
And lastly, even retail brokers don’t give you anything for free and all of them (as far as I know) will charge (or credit) carry interest, as this is an additional revenue stream for them aside from the bid/ask spread.