Trend System on Other Assets

This post is a continuation of a series on the multi-model trend system which I have developed. In the past weeks we looked at different applications of my trend model. First, I looked at several backtests on the S&P 500. I then applied the exact same trend system on other major equity indices: the Nikkei 225, NASDAQ Composite, and CAC 40.

The following links cover these backtests and some other information about the system I developed.

Long and Short Trend Systems

Long/Flat Equity Trend Systems

Long Trend System on Other Equity Indices

Reminding readers... I have developed two trend systems which have many similarities in their structure. The major difference is the time examined. The shorter duration system uses a lookback period of trends that are identified at two weeks through three months. The longer duration system sees trends forming at three months through twelve months. I've called them the short-term trend system and long-term trend system.

In this post I will apply the exact same systems on other assets. This includes currencies (EUR/USD and JPY/USD), interest rates (10 Year Treasury Yield), and alternative commodities (gold and Bitcoin). The currencies and interest rates will be using the long-term trend strategy that goes both long and short the asset, betting on either side of the trade based on the trend.

Gold will be using the long-term system in a long/flat format (betting with the trend fully or partially, or being in cash). Since Bitcoin is a newer asset that is highly volatile and difficult to short, I modeled it long/flat with the short-term trend system.

Note: All data is calculated in nominal terms and is not adjusted for local interest rates or inflation. The objective is to demonstrate the behaviour of my trend system across a range of markets. These are simulated results.

Euro/U.S. Dollar Currency Pair

Since the Euro was created as the new pan-European currency it has become the dominant alternative currency to the U.S. dollar. It is also the biggest currency pair trade with a massive number of futures contracts trading hands each day.

The EUR/USD futures contract is a great place to start for people aspiring to be currency traders. One contract controls delivery of 125,000 Euros. This contract requires a margin (deposit) of just $2,000 to hold.

Understand and trade based on the contract value, not the margin requirement! As you can see, the leverage factor is enormous.

The following charts backtest my long-term trend system using absolutely no leverage. This simulates the underlying return of the strategy. Most traders would use some leverage in practice.

Total Performance (Growth of $100)

Credit: TheRichMoose.com, Quandl.com.
Right-click to expand image.

Drawdowns

Credit: TheRichMoose.com, Quandl.com.
Right-click to expand image.

3-Year Rolling Returns

Credit: TheRichMoose.com, Quandl.com.
Right-click to expand image.

If an investor were to buy 100 Euros in 2000 when this backtest started, they would have paid about $102. Today it would be worth about $112, translating to an approximate 10 percent gain to the investor over the backtest period.

With my trend system, that same $100 would have grown to be worth about $168. That implies a gain 5x larger than just holding the Euro.

The Euro/U.S. Dollar is a very stable currency pair to trade and a great place for investors to start with currency trading. The largest drawdown in my trend system was about 18 percent. Many traders can comfortably use some leverage in this asset to increase their exposure on the trade.

Japanese Yen/U.S. Dollar Currency Pair

The second largest currency pair trade is the Japanese yen against the U.S. dollar. As with the EUR/USD contract, this is a major futures contract that trades over 100,000 contracts each day.

A single contract controls delivery of 12.5 million yen worth approximately $115,000 at current rates. The margin requirement for the contract is just $1,800.

The following backtest and charts use no leverage to simulate the underlying return of the strategy. It is a long/short trading strategy as it is extremely easy to short currencies and futures contracts.

Total Performance (Growth of $100)

Credit: TheRichMoose.com, Quandl.com.
Right-click to expand image.

Drawdowns

Credit: TheRichMoose.com, Quandl.com.
Right-click to expand image.

3-Year Rolling Returns

Credit: TheRichMoose.com, Quandl.com.
Right-click to expand image.

Futures contracts on the JPY/USD currency trade have existed for a long time. My data goes back to 1977 and covers a range of market conditions for the Japanese and U.S. economy. If an investor were to buy $100 worth of yen in 1977, that would have grown to be worth about $240 today. That's a 140 percent gain.

With my trading system, $100 would have grown to nearly $600 in that same time period.

As with the Euro, the yen is a pretty stable currency in general. However, we do see quite large moves from time to time. Drawdowns with my system peaked at 36 percent. We also see a relatively long period of declining returns from the late-1990s to 2011. Since then new highs were made with the system.

Looking at the 3-year annualized rolling returns, we can see the sharp difference between the return characteristics of currencies compared with stocks when applying a momentum strategy to both markets. Returns on the yen do not see significant negative returns when smoothed, but the declines are grouped together into several years of poor performance.

Interest Rate on 10 Year Treasury

Trading interest rates can offer investors a return stream that is very unique, different from bond investing, but yet relatively stable. The interest rate often take the inverse of bond returns in the shorter term. When interest rates go up, bond values drop; when interest rates drop, bonds jump up.

Interest rates can be trading via futures contracts. I would not recommend short-selling a bond ETF. Futures allow you to trade more closely in line with interest rates in a liquid market. A single contract controls $100,000 of Treasury notes.

As with currency pairs, with futures it is easy to go long or short the contract (buy or sell) at a very low margin requirement of just $1,050 on the 10-Year Note Contract.

Total Performance (Growth of $100)

Credit: TheRichMoose.com, Yahoo Finance.
Right-click to expand image.

Drawdowns

Credit: TheRichMoose.com, Yahoo Finance.
Right-click to expand image.

3-Year Rolling Returns

Credit: TheRichMoose.com, Yahoo Finance.
Right-click to expand image.

Yields on 10 Year Treasuries often move slow; a long-term trend system works particularly well. If an investor put $100 into this strategy in 1963, it would have grown at a pretty steady rate to nearly $4,000 today.

Drawdowns reached about 40 percent, but in the backtest the recoveries were very quick compared with many other assets we looked at. The drawdowns seem to have gotten bigger in recent years. This may be indicative of a shift in the market characteristics.

Gold

In this backtest I used my long-term trend strategy on gold, buying and selling in U.S. dollars. Gold has long been a favourite alternative asset for many investors. I would describe gold as a sort of crisis asset. It often does well when many other markets do not. Being the classic "hard asset", it is particularly a good asset to hold when trust in currencies fades.

Although it isn't necessarily difficult to short-sell commodities via futures contracts, I don't think it makes as much sense as short-selling currency pairs or interest rates. With commodities short-sellers have the headwinds of a long-term upwards bias due to inflation and depletion of resources.

Gold can be purchased with ETFs such as GLD or IAU and with futures contracts. A single gold contract controls delivery of 100 ounces of gold. It is a highly liquid contract with roughly 270,000 contracts traded daily.

In this backtest I compared my long-term trend strategy against simply buying gold and holding it throughout the period.

Total Performance

Credit: TheRichMoose.com, Quandl.com, FRED-Federal Reserve St. Louis.
Right-click to expand image.

Drawdowns

Credit: TheRichMoose.com, Quandl.com, FRED-Federal Reserve St. Louis.
Right-click to expand image.

3-Year Rolling Returns

Credit: TheRichMoose.com, Quandl.com, FRED-Federal Reserve St. Louis.
Right-click to expand image.

Many investors see gold as a safe asset that is linked to inflation and should always go up. This isn't the case. Since the U.S. dollar was delinked from gold, the yellow metal has had periods of high returns and long periods of negative returns.

Even during the inflation period of the early-1980s, gold actually dropped in price. If you bought gold in 1979, you didn't see a profit until 2008. Instead of being safe, gold took a massive 70 percent drawdown over 30 years.

Using a trend system, we substantially reduced the severity of the drawdowns. We also saw nice gains when gold moved up. The overall return for the trend system was only about 4x higher, the returns were more stable and allowed long periods where an investor could do other things with their money (held in cash in this scenario).

Bitcoin

Bitcoin is a newer asset that still carries many questions. While certainly interesting, its viability as a usable currency is not there in my view. I also have a hard time seeing Bitcoin as a safe asset with gold characteristics.

Since Bitcoin was spawned, many other cryptocurrencies with improved technology and capabilities have come to the market. One day we could see some type of decentralized cryptocurrency used in daily life. We are definitely some time away from that yet.

Right now I would say that Bitcoin in simply another asset to speculate with. It is not a buy-and-hold item. It is not where anyone should put a substantial amount of their wealth. There are issues with security, storage, scalability, among other problems.

Thankfully it is getting better. There is an increasingly popular Bitcoin ETN on the market trading with the ticker GBTC. This makes Bitcoin a bit more accessible and possibly a bit more secure to hold, albeit with regulated third-party risk.

Bitcoin is a new, volatile asset. It is also not easy to short-sell. For this reason I applied my short-term trend strategy in a long/flat format, compared to simply buying and holding a single Bitcoin.

Total Performance

Credit: TheRichMoose.com, Yahoo Finance.
Right-click to expand image.

Drawdowns

Credit: TheRichMoose.com, Yahoo Finance.
Right-click to expand image.

Annualized Rolling Returns

Credit: TheRichMoose.com, Yahoo Finance.
Right-click to expand image.

Bitcoin is a truly crazy asset. Since 2010, it has seen numerous 90 percent drawdowns. Bitcoin also saw enormous annualized returns. As the market is maturing, the swings are getting less severe.

My trend system cut drawdowns on Bitcoin nearly in half and, as usual, nicely smoothed returns for an otherwise rough asset. By dancing out of the worst of the drawdowns, the trend system would have returned over 3x more to the investor.

For those interested where Bitcoin fits into the portfolio, this article recommended investors have no more than about 2 percent of their portfolio in Bitcoin. Many investors should probably have none.

Is Bitcoin in the Optimal Portfolio? - Of Dollars and Data

Personally I believe you should risk no more money in cryptocurrencies than in any other asset. The rule of thumb is to risk about 1 to 3 percent of your assets per trade, depending on your level of aggression. Since Bitcoin and cryptocurrencies have huge drops in value and could realistically become worthless, that means you should have a maximum of 1 - 3 percent of your assets in cryptocurrencies.

I currently don't hold any cryptocurrency but I am watching some of the developments on the newer third-generation crypto projects with interest.

Summary

A good investment system is not limited to a single market. It is applicable and effective across a range of markets experiencing very different market conditions.

I will continue to update readers on the momentum factors of the markets I am personally invested in. In the meantime, I will be working on my systems trying to improve the systems and make them easier to invest with for my own purposes.

I'm interested in hearing your questions about investing or personal finance and ideas for blog posts. Please leave a comment or send me an email at: richmooseblog @ gmail . com.

Comments & Questions

All comments are moderated before being posted for public viewing. Please don't send in multiple comments if yours doesn't appear right away. It can take up to 24 hours before comments are posted.

Comments containing links or "trolling" will not be posted. Comments with profane language or those which reveal personal information will be edited by moderator.

6 Replies to “Trend System on Other Assets”

  1. Thanks for the article Daren.

    With your long-term system in a long/flat format you mention you are either fully invested in the asset, fully invested in cash or partially invested in the asset. Similar to DM, except with DM you are 100% in the chosen asset, bonds or equities. Can you explain how the partial works here – do you have a certain number of thresholds? This reminds me a bit of the article you discussed a while back that described DM with hurdles where you phased your exposure out or in, in increments.

    Also, what are you plans for your models – do you intend to share them at some point or sell a subscription? Finally, how similar are they to the Newfound U.S. Trend Equity Index?

    1. Daren (Editor) says:

      Currently the model puts you into an allocation between 0 – 10 as a part of a full allocation. This would be appropriate if you purchased standard ETFs and had a larger allocation as a percentage of your portfolio. Right now the score for SPY is +9, so your portfolio may be 90 percent in SPY and 10 percent in BSV or BND or whatever.
      I’ve been working on the model and I think it’s best to use a scale between 0 – 5 or even 0 – 4 when using something like my Leveraged Barbell Portfolio where we would use LEAPS options or leveraged ETFs. In a 0 – 5 scale, +5 would be full allocation, +3 would be 60 percent, +1 would be 20 percent, etc. Jumping by 20 or 25 percent allocations is just as effective as incremental 10 percent or even 1 percent allocations.
      I personally like using short-term bonds. So when I say cash, hear short-term bonds. Depending on the style of investing you use, that cash could effectively put in short-term bonds, a broad bond fund, or even different investment opportunities that are relatively liquid like long/short currencies.
      This style is quite different from Dual Momentum in many ways. DM picks just one market between three and invests completely in the strongest market based on a very specific measure of momentum. There is a dynamic method I’ve been working on to invest in a DM format using my much more comprehensive momentum model. It wouldn’t scale in and out though, I think that gets very tricky with way too many trades.
      I would say my trend models are best suited to a style of investing like the Leveraged Barbell Portfolio. For example, if you invested only in U.S. stocks, emerging markets, and long-term bonds, you may allocate something like a maximum 15 percent to each asset in a 3x leveraged ETF. The remaining would be in short-term bonds. If the current score (on a 0 – 5 scale) for U.S. stocks is 5, EM is 3, LT Bonds is 3, your portfolio might be about 15 percent UPRO, 9 percent EDC, 9 percent TMF, and the rest in BSV.
      My models are similar, but different from the Newfound U.S. Trend Index. My model has addition methods of measuring trend which are academically supported. It also tilts slightly more to the shorter-term, going as low as about 3 month trends. This is to account for a broader range of market conditions that we can see in history. The U.S. is one of the only country markets that is most effectively invested with very long-term trends (10-12 months) when measured right now. That was different if you had measured the U.S. market in 1985. The U.S. is also the most successful market of the 20th century. This makes it tempting to be affected by survivorship bias, hindsight bias, and anchoring when it comes to investing. When making comparisons, it can be difficult on the surface to see the different between any decent trend measurement. Newfound’s model and my model are also similar to using a 10-month SMA, or a 200-day SMA, or 12-month time-series momentum, or any other single measurement of longer-term momentum. The differences come when looking under the hood and looking across markets.

      I’m working on several different options to share my trend models. One of those is subscription based, maybe offering several markets for free. The challenge is figuring out the format and how to disperse the information effectively.

  2. Yes I thought this made sense to use with a leveraged barbell portfolio. Wondering though in your example with 4 holdings and a 0-5 scale for 3 of those could result in quite a lot of re-balancing. To keep things simple I am currently using the 15% UPRO and 85% ISTB method you discussed with the re-balancing occurring during a crossover of the 13-week SMA. However, I was thinking that per a momentum signal it would make sense to increase my leveraged equities exposure, likely to a maximum of 40%. Just grappling with the most effective method of doing so. Using the Newfound % of positive trend signals was thinking something along the lines of 0-50% positive = 15% leverage; 50-75% = 30% leverage; and 75-100% = 40% leverage. And re-balance whenever one of those thresholds is crossed. Do you think that makes sense?

    Thanks.

    1. Daren (Editor) says:

      I think that makes sense. But you need to be prepared for significantly increased volatility in your portfolio. With 15/85 UPRO/ISTB, you are getting something close to standard 60/40 portfolio performance with less volatility. (The outsized performance over the long-term comes from less exposure to stocks in big downtrends and compounding gains with leverage in slow uptrends.) When you ratchet that up to 40/60 UPRO/ISTB things can move a lot in a short period of time.
      The 0-5 scale doesn’t change things significantly as far as re-balancing goes. Looking back to 1950 there’s a pretty consistent pattern. For a month or two you will be making small adjustments every week or two. Then for many months you will be at full allocation (5) or no allocation (0) with no adjustments. This cycle repeats itself. Sometimes several months into an uptrend you will get a signal from 5 to 4 and back to 5 again, a good time to take profits and re-align your allocation.

    2. Daren (Editor) says:

      Forgot to mention… With 4 holdings rather than 2, you are spreading your bets as to which market will perform the best. It’s easy to look at the past ten years and put your money on UPRO. Heck, use TQQQ or TECL for that matter. But that assumes the future will be exactly like the past ten years. Things never work out like that.
      If you were running a big trend hedge fund, you will find that it is best to get exposure to as many markets as possible. It increases your odds of catching a market that runs and lowers your odds of getting wiped out on a single market or two. This is because markets behave strangely. Even though heating oil and crude oil are correlated around 0.95, there have been instances where heating oil goes on a massive run while crude sits flat. If you had just picked crude as a proxy for all energy, you would have missed out. There are many examples of this across all market categories. Just recently it was palladium relative to other precious metals.
      As individual investors with a relatively small amount of money, we don’t have the luxury of spreading across dozens of markets effectively in a way that makes sense. But we can definitely catch 4 or 8 or whatever your portfolio size allows.

  3. Appreciate your responses – very insightful.

    Regarding your second post I did fail to mention in my example that my allocation to leveraged equities would be determined in a DM fashion where I would look at broad indices for US, Developed and Emerging and choose the leveraged ETF for whichever had the strongest momentum over the previous 12 months. With that approach I am not as diversified as your 3 + BSV example, however, I would always be either holding, or rotating into what had the best momentum on the equities side.

Leave a Reply

three + five =