I've been looking at the trending potential of currencies recently, noting that they are more prone to trend following than mean reversion type pull backs.
In my test I performed a simple analysis looking at the performance of the DVB indicator. In both mean reversion and trend following mode. As a brief recap, the DVB is a short term overbought/ oversold indicator which ranges between 1 and 0. It performs a similar function to the RSI set to 2 periods. The general idea for markets which pull back a lot is to sell on DVB readings above 0.5 and buy on readings below 0.5. A trend following method would do the opposite (Buy on strong readings and vice versa).
ETFs are hugely popular stateside and are gaining interest in the UK. Unfortunately brokerage fees for UK stocks and UK listed ETFs are ridiculously expensive compared to US ETFs available through discount brokerage firms like Interactive Brokers.
So I also wanted to test currencies using the spot rate.
I looked at the EUR/ USD exchange rate.
ETF: FXE (US listed)
Spot rate: EUR/ USD
I then ran a simple strategy test as follows:
Mean Reversion (MR): Buy on DVB reading below 0.5. Sell above 0.5
Trend following (TF): Buy on DVB reading above 0.5. Sell below 0.5
Data back to 2010 only. No commission etc. FXE data dividend adjusted.
In theory the ETF and the spot rate should respond in a very similar manner, but do they? In a previous post I showed how the FTSE 100 index reacts in to MR strategies in a completely different way to its US listed ETF.
I thought that FXE and EURUSD would surely be closely aligned….
First here's the performance of the simple strategy on the FXE ETF. MR has a slight edge, but there's no much to shout about really.
I've no idea. It could be the non 24 hour nature of the ETF skewing the DVB readings. Perhaps the dividend adjusted data throws the interest rate differential out of whack. Maybe people play ETFS different around dividend time. Maybe I'm missing something, but the takeaway from this is that if you intend to trade certain market, run any tests on a vehicle as close to the actual market you'll be trading as possible. E.g. If trading ETFs, don't assume the spot or index data data will be the same.