Market timing and style rotation

Discussion in 'Trading' started by Junkou, Apr 13, 2022.

  1. Junkou

    Junkou

    I will answer the 3 questions to tell the story about market timing and style rotation:
    • Question 1: can you predict the short-term direction of Apple share price
    It's very difficult if not impossible.

    If we write in a formula: d=f(x). There is only one variable, which has no boundaries and is free to move in any directions. You may say there are moving average lines that can be boundaries. However, these lines are not binding and the price is not necessary to move to those lines. Furthermore, those lines are also moving.
    • Question 2: can you predict whether Apple share price will outperform the S&P 500 Index in the short term
    Short answer: it's possible even though still possible

    We now have two variables in the formula: d=f(x-y). With the S&P 500 Index as the reference, the range that Apple share price can move from the S&P 500 Index is relative stable and predictable. We can calculate the current standard deviation from its long-term mean. If it's in the middle, we can't predict the direction. If the current position is over 2 standard deviation from the long-term mean, we can say Apple share price is likely to underperform the index. On the other hand, if the current position is 2 standard deviation below the long-term mean, we can say Apple share price is likely to outperform the index.

    • Question 3: Can you predict whether NASDAQ 100 INDEX will outperform the Russell 2000 Index in the short term
    It's possible and relative easy.

    Indexes in the US market can be classified as different styles. What is a style? A style is the common feature that drives the performance of a group stocks. The widely used styles are value and growth, and large and small. For example, the NASDAQ 100 Index is largecap stocks with growth oriented, and can be classified as growth and large. The Russell 2000 Index is small size and value.

    How to model the relative strength of two styles then?

    There are several steps:

    1. Calculate the relative performance of two styles. In this example, we can simple use the price ratio between the NASDAQ 100 INDEX and the Russell 2000 Index
    2. Apply standard pairs trading model to calculate the current standard deviation
    3. If the position is in the middle, use statistic models to determine whether there is a trend. Machine learning models can be used to estimate the thresholds
    4. If the position is at the upper bound or lower bound, apply mean reversion models to determine the reversion points. Again machine learning models can be used to estimate and optimize the thresholds
    5. If the position is in other places, use rules to control risks
    Many academic researchers use daily data or even monthly data in their study. That's a mistake. Price-based models should include intraday pricing which contains more information.

    How can we use the prediction from style rotation?

    Long-only investors can change holdings with index-based ETFs accordingly.

    A long/short portfolio can also be set up to achieve returns with very low correlation to the market.

    Happy to share more details if you like this topic.
     
  2. SunTrader

    SunTrader

    Baron needs to start a Theory (leave Trading one for ... Trading) sub-forum. Except 3/4 of threads would eventually wind up there lol.
     
    murray t turtle and Baron like this.
  3. Junkou

    Junkou

    Time for growth to shine: rotating into growth ETFs such as IWF and QQQ. If hedge, using value and small size ETFs such as DIA and IWF.
     
  4. Overnight

    Overnight

    Growth is about to collapse into the abyss. Growth and rate hikes do not mix, no sir. Keep focusing on value.
     
  5. Junkou

    Junkou

    I'm talking about the short-term performance. Value is still better in the long-term during a rising rates environment.

    We will see in next few days.
     
  6. Overnight

    Overnight

    Yer high, man.
     
    David's faith likes this.
  7. mikeriley

    mikeriley

    Good grief, that's a lot of calculations. Here's an easier way, just
    follow trend action as price bounces back and forth
    within a beautiful range. Like Nasdaq has done all week (Picture below)


    range.png
     
  8. nitrene

    nitrene

    IWM used to simply be a high beta trade much like QQQ, but I think the last 3-4 years it is really just a leveraged value play. Recently I noticed that DIA & IWM do move in tandem and QQQ is the reverse of DIA especially when rates go higher. The Dow was up most of the day while QQQ was down 1%+ all day long. There must be algos running this dichotomy based on the 10-year rates.
     
  9. Junkou

    Junkou

    Exactly! The Russell 2000 Index has become a value index with small size tilt. It's highly correlated with DIA in the last few years. The Dow Jones index is widely seemed as a proxy to the value, similar to IWD but with futures. Institutional investors tend to use e-mini futures to get quick exposure to either value and growth, the widest followed styles.

    Following is an example from the process outlined in the beginning:
     
    easymon1 and nitrene like this.
  10. nitrene

    nitrene

    I like IWM since there are leveraged versions of it both long & short (TNA & TZA) as well as futures which is even higher leverage depending on your broker. IWN is probably the ETF to use for even better alignment to the Value sectors but it doesn't have the liquidity.
     
    #10     Apr 15, 2022