up up up

That's interesting. Theoretically, the market should still be efficient without short selling, but I wonder if you'd see something similar to what you see with closed-end funds, namely shares can trade above a fund's NAV. Though that is caused by a different mechanism.

My hunch is that you're on the right path with momentum and/or mean reversion. I would expect that a market without brakes would have more momentum, and that the trends would reverse rapidly when the momentum ends.

Are you allowed to trade options? That would allow you to get short exposure that way -- though it would also allow anyone to get short exposure, likely getting rid of the opportunity.

HI James,

Thanks for your reply. The options market still very small. So i think that's not really liquid enough to trade.

Cheers

Hi Welly,

Technical Analysis - That's a hard one. If it works, it works. If not, it doesn't. I think you could backtest if technical analysis works, but be wary of the economic theory around it. I would recommend you test it across multiple markets and get an understanding why it works. Something like a 10-month Simple Moving Average can be very useful to determine the general trend of the market.

Momentum is real! It works on a diverse number of markets, but it's not intuitive. You have to manage risk very carefully with momentum, and ideally combine it with some value factor. Otherwise, you might see dangerous drawdowns you might not like!

Mean reversion - Definitely! It won't be as clean as a Long-short mean reversion, but you could easily buy high quality companies that dip due to CAPM or some other reasoning. Just be careful to filter for some quality or value factor. From my experience, if you KNOW the company is a really good one, buying the mean reversion gives you that confidence.

@Chen Peng,

So basically no place for automation then. we need to treat that based on fundamental analysis + technical analysis.

Is there any way we can find the appropriate automated trading model for this kind of market ?

Cheers

Actually, the difficulty in trading momentum and technical analysis is the exact reason for automation (and it's very likely profitable)!

I find there are numerous ways to trade momentum. Some of those I have used are Price Momentum (12-1m), Relative Momentum (Dual Momentum investing by Gary Antonacci). I emphasized fundamental factors greatly because it compliments momentum extremely well since they are both uncorrelated.

An example could be:
- Filter the universe of stocks by the highest momentum and pick the top half
- Sort this new universe by some fundamental factor like return on equity
- Size positions based on the differences in momentum and quality

I have a long-only strategy myself that seems to work based on these principles

@Cheng Peng,

How to measure momentum, did you use technical moving average for that or you use some statistical test ?

Cheers

Hi Welly,

There are many ways to measure momentum. My personal favorite is the 12-1month price momentum. You can find how it's calculated here: https://www.quant-investing.com/news/2016/10/13/twelve-months-minus-one-month-momentum-added-to-the-screener-(12-1-month-momentum)

Cheers

Cheng Peng,

Can we use something like Hurst Exponent for measuring momentum ? Or that's only for mean reversion only ?

We could use the Hurst Exponent as a factor to determine if a security is mean-reverting, geometric motion or trending based on the results.

I believe Hurst Exponents that have results > 0.5 are generally accepted as trending. I personally haven't found too much success in using it this way, and it seems to be more popular in stat-arb. Try and test it out!

I design a lot of long only baskets. This algorithm will rank them by their 26*4 alpha score, and chooses the stocks with the highest momentum over the last quarter. It then buys them as they come in, and optimizes them based on their alpha ranking. I have a few others, if you want to see them. It still needs some work on the leverage because it is only buying half the leverage since the leverage is on 2.

@eric.

Nice posting. Thanks for the a algorithm.

Yes i would like to see more.

Cheers