Page 1 of 1

### Backtesting and Accounting for Carry

Posted: Tue Jul 01, 2003 11:12 am
I have not traded forex yet myself but intend to do so in the near future.

Getting historical data is possible, however, the data appears only to have price information. There is no record of the cost/profit of carry that would be incurred when taking a position due to the interest rate differential.

This cost/proift appears to be significant at times. I have a few questions:
1. Has anyone here tested forex while accounting for the interest rate differential? Is it as significant as it appears it might be?
2. Does anyone know a good source for historical interest rates as used by forex brokers to determine the cost of carry?
3. Is this something we should be spending our time on?
Thanks,

c.f.

Posted: Tue Jul 01, 2003 1:30 pm

I just spoke to FXCM and they say that the interest rate differential is almost not worth bothering about.

The following is a simple example of how interest rollover is calculated:

No. of Lots x No. of Units per lot x Yearly Interest Rate Differential / 360 x No. of Days

Example:

Transaction: Sell 2 lots of USD/JPY on Monday and settle on the next day
Lot Value: USD 100,000 or JPY12,200,000
Opening Price: 122.00
Yearly Interest Rate Differential, USD 1.25% - JP .0% = 1.25%
Calculation: USD 100,000 x 2 x (-1.25%/360) x 1 = -6.94

Example:

Transaction: Buy 2 contracts of GBP/USD on Monday and settle on the next day
Contract Value: GBP100,000
Opening Price: 1.5600
Yearly Interest Rate Differential, GB 3.75% - US 1.25% = 2.50%
Calculation: GBP100,000 x 2 x (2.50%/360) x 1 = 13.88

These examples are approximations and only serve as a guide. Real interest rate charges will differ from these calculations because like the FX market, these interest rates are also quoted as a spread.

Posted: Tue Jul 08, 2003 11:51 pm
Vince,

If you trade longer term this actually does make a difference. The current interest charge for a 1 lot GBP/USD spot position is roughly \$10 a day -- for a 2 lot position held for 4 months this adds up to 10 X 2 X 120 = \$2400. Pretty significant extra profit or slippage to take into account.

As one trader mentioned in a recent post, it has been a profitable strategy for many traders to take "carry trades" -- simply putting on a position that earns the carry over time. Because frequently the currencies of countries with higher interest rates tend to appreciate over time against those of countries with lower interest rates as capital flows out of the low interest rate (ie, unattractive for investment) countries and into the higher interest rate (more attractive) countries, this simple strategy may be profitable in and of itself and, guess what, you get to earn the carry too.

Posted: Wed Jul 09, 2003 3:34 am
an illustration : MXN in blue (without carry) VS MXN future (with carry) in red.

Posted: Wed Jul 09, 2003 8:34 am
Nice chart Moodaeng -- this is an example where my tenet, that higher carry currencies generally tend to appreciate against lower carry ones, didn't hold up. The Peso lost value over the period shown while Mexican interest rates were higher than those in the US. BUT, clearly if you'd been long the peso the interest paid would have outweighed the depreciation and you'd have been net profitable.

I believe there are some serious benefits to using a naive diversification strategy to trading currencies -- a combination of trendfollowing, carry trading, and perhaps short options may be able to produce excellent risk-adjusted returns.

-- Neal Stevens

Posted: Wed Jul 09, 2003 1:21 pm
That looks great, Moodange, and I agree, the carry becomes more important with emerging market currencies.
There is a realistic mean reversion when you trade G-7 that doesn't exist when you trade say BRL.
Plus, the further out you go in time the more it becomes a pure fixed income instrument, especially when you use forward/forwards. It = the currency contract(s).

Jester

Posted: Sat Aug 02, 2003 12:07 am
Actually, the general consensus in the academic community seems to be that the "forward bias" that brings high interest rate currencies to appreciate against low interest rate currencies only applies with respect to currencies of developed countries.

There's quite a bit of econometric literature on this that you can find for free on the internet -- just do a google search for "forward bias trading" or "forward bias puzzle". FX Concepts, which trades more than \$6 Billion (yes, that's a B) using a mix of trendfollowing and carry trading and has achieved a sharpe ratio of 1.8 doing so, has quite a bit of material on their site on this too.

Posted: Tue Aug 05, 2003 10:36 am
Bollinger,
I'm not really understanding your point. The whole idea of trading an emerging market currency is to trade it against a stable one. You could trade emerging market crosses, but if say, Brazil is blowing up, why trade it against ARS?
GR

Posted: Wed Aug 06, 2003 2:52 am
Ghostrider:

(1) There is a recognized phenomenon that among two developed countries the currency of the country with higher interest rates will have a tendency to appreciate against the currency of the country with lower interest rates;

(2) This phenomenon is not recognized to occur with respect to currencies of emerging countries.

### Rollover

Posted: Tue Sep 16, 2003 2:42 pm
I am trading a turtle-style trend-following system in Forex and have been for about a year now. I have done well as a dollar bear. As a long-term trendfollower I DO NOT ignore the rollover. In fact, I use it to my advantage. About 13% of my profits have come from the rollover. I only take spot trades when the rollover is in my favor.

It is my assumption (one that I am testing) that smooth long term trends occur in markets that have an interest rate advantage. If a market lacks the interest advantage it should move in a choppy trend driven by intra-day traders. Of course there are other reasons for a market to trend besides interest rates, but an interest rate advantage should help smooth a trend.

Posted: Sat Sep 20, 2003 1:29 am
Bollinger wrote:If you trade longer term this actually does make a difference. The current interest charge for a 1 lot GBP/USD spot position is roughly \$10 a day -- for a 2 lot position held for 4 months this adds up to 10 X 2 X 120 = \$2400. Pretty significant extra profit or slippage to take into account.
Another way to look at the significance of such an insignificant number is as follows:

Let's say you have developed a Long Term system that trades the GBP and it nets out \$300 average per trade. This is the net, so the system has already debited all expenses, except for the carry. This being a LT system, let's say it holds the trade for 30 days.... with these numbers the system is only making \$10 per day per contract. When you have to account for an additional expense like carry, you are no longer looking at \$300 net per trade, you are now looking at break even.

Vince asked FXCM and it is in their best interest to say that the cost of carry is "almost not worth bothering with." They want your business and they are good at sales not at trading.

This also points to the importance of knowing who you are taking counsel from.
• If you ask a drinker for advice, he'll say "let's have a few drinks and forgot about it."
• If you ask a religious person they will say, "let us pray about it."
• If you ask a broker about something that will impact your bottom line and not his, he will say "almost not worth bothering with."
• If you ask a trader, he will say.. "It does matter! I wouldn't trade against it."

Gordon

### Re: Backtesting and Accounting for Carry

Posted: Mon Oct 06, 2003 4:33 am

### Re: Backtesting and Accounting for Carry

Posted: Mon Nov 24, 2003 10:20 am
Forum Mgmnt wrote:
1. Has anyone here tested forex while accounting for the interest rate differential? Is it as significant as it appears it might be?
2. Does anyone know a good source for historical interest rates as used by forex brokers to determine the cost of carry?
3. Is this something we should be spending our time on?
Forex actually brought me to mechanical systems trading in the first place, some 1 1/2 years ago. But getting reliable historical price data seems to be impossible without access to professional (and very expensive) systems such as bloomberg or Reuters. The same seems to hold for interest rates.

In any case, obviously lots of additional coding/computing would be necessary to determine cost of carry from an available interest rate history. That's way beyond my own capabilities. However, it's probably a piece of cake for most of the traders in this remarkable forum.

Anyway, here's a link I found in my favourites (from a time when I still thought I could master FX ) which might be helpful: http://www.tmpages.com/tmp55.htm

### Forex data

Posted: Tue Feb 17, 2004 2:33 am
Wojo,

Have you tried CRB (Commodity Research Bureau) for FOREX data. I use it for system testing. It covers most international markets including cash and contract markets and futures. Looking at my most recent bill it is 20USD per month.

I can't vouch for the reliability as it uses multiple sources from around the world but it's a useful scratch pad and starting point.

Cheers,

Stephen

Posted: Wed Dec 22, 2004 12:06 am
Let's assume that you are following a trend in the Pound. You have a choice of trading your Pound position against \$, Euro, Yen, Swiss, etc. Your preference, from a carry perspective, would be to trade long Pound vs. Yen, since that is where you would get the largest carry differential, right?

I guess this would be a good way to help select the optimal currency pair to trade.

Posted: Wed Dec 29, 2004 9:51 pm
Carry is important, but is not more important than being long/short the strongest/weakest horse out of the gate. I find many times the large carry trade agrees with the best trade.

Posted: Fri Dec 31, 2004 5:28 pm
[quote="Ghostrider"]Carry is important, but is not more important than being long/short the strongest/weakest horse out of the gate. I find many times the large carry trade agrees with the best trade.

Self-reinforcing trend perhaps? Would be interesting to backtest, no?

Posted: Sat Jan 01, 2005 12:28 am
Nice point, and good question.
My tests show that price breaks usually preceed carry. The trick is to watch when the institutions pile into the carry trade, they are the ones that really push the move.