Question about Aggressive/Defensive Book

Research
1

Hi All,

Before I head down a dead end and waste a lot of time, I am wondering if it is possible to create a book with one portfolio that is aggressively long and another portfolio that chooses between several defensive vehicles?

More specifically, I've built a pretty decent simulation that chooses the optimum two stocks from the Mag 7 based on momentum. It produces a decent 30% return using a hedge component, based on a composite agg series using breadth, volatility, and some secret sauce. The sim uses TLT as the defensive asset.

TLT worked pretty well until (primarily) the 2022 market decline. Of course, that downturn wasn't caused by an economic recession; it was caused by the Fed raising rates to battle inflation. When the Fed is raising rates, bonds are never going to work as a hedge. If I simulate using an exposure list that goes to cash during 2022, the performance jumps to 50% annualized.

If I was running it live in 2022, I would have overruled it an gone to cash instead of TLT. However, I want to run a 100% model-driven strategy, with no manual decision-making involved.

So I am wondering if I can build a defensive port that uses a ranking system to select the best asset out of perhaps, TLT, BIL, SH, or cash (or others?). The defensive port would use the same risk-control agg series for in-market signals (reversed of course).

I would then put them together in a book, and hopefully be able to have a portfolio that selects the optimum asset regardless of whether market conditions are positive or negative.

However, before I head down the very time-consuming road of building all that, I am wondering if anyone else uses this idea and been successful with it?

Sorry if this is a stupid question, but I'm relatively new here and this will be my first book portfolio. Thanks in advance.

Chawasri

2

Chawasri,

I want to help you but I don't use books. What you are asking is not a simple P123 user question. I don't think many existing users (including customer service hotline) can help you. Maybe hemmerling (Kurtis) can give you some assistance.

I use the download function to export P123 data and work on the spreadsheet. (not something fully automated from P123 that you seem to expect.)

For the risk-off asset, you can try adding GLD as a defensive asset.

Regards
James

3

Yes, I intended to put GLD in my defensive list, but forgot it in the moment. Thanks. By the way, even GLD was a loser in 2022. The only proper investment for most of that year was a money market fund (which is what I did).

Nevertheless, I'm sure there are users putting the Books feature to use. I hope to hear from them if they have taken my proposed approach.

4

I have a couple of questions before delving into this.

  1. What is it about a book that will be different from using a constant hedge, as you're doing now?

  2. Using the Mag 7 in a simulation is a classic example of look-ahead bias (or selection bias). To avoid this, you should probably run a simulation that chooses among the largest seven US stocks by market cap at any one time. You can do this by creating a universe with one rule: FOrder("MktCap") <= 7.

3 Likes
5

Hi Yuval,

  1. Thanks for the reply. I thought I explained pretty well (but I guess not) that there are times when the constant TLT hedge is not optimum. As I explained, TLT lost significant money in 2022 when interest rates were increasing. A book will have a long strategy and a defensive strategy, choosing the optimum defensive assets for conditions when the risk-control signals for that component to kick in. In 2022 (for example), most everything was declining, so cash would have been the optimum asset.

  2. I have other strategies that are more diverse, however, I have been running a Mag 7 strategy for the last 12 years and it has performed spectacularly while using proper risk controls and defensive assets. I want to get that model set up on this platform, if possible.

6

Yes, a book would be the way to go in this case. I think it suits your purposes fine.

The Mag 7 didn't exist until 2023. That's when Michael Harnett came up with this particular list of seven companies.

7

Previously they were called the FANG stocks, FANG+, FANGMAN, and other variations. It didn't matter the acronym or name, they were the largest, most successful tech stocks that dominated their industry niche. And maybe more importantly, they are constantly in the news and discussed on social media, generating interest and activity that fuels their price.

But I guess I was asking specifically if anyone had created a book that uses the same series for control on 'in-market' conditions with both a long and defensive book components.

Can the long portfolio use, say, a >=0 series signal to be invested long, and the defensive portfolio use the same 'in-market' agg series – reversed (<0) – to select the optimum defensive position? Will the system work together with no weeks out of the market because of an internal lag or other issue?

Maybe the only thing I can do is dig in and try it by creating my first book, but I was hoping to get a sign that someone had tried this already before diving in and finding out it won't work after a couple of weeks are invested. Thanks much.

8

Hi Yuval,

I decided to try your approach using FOrder("MktCap") <= 7 as the universe definition on the Nasdaq 100. However, the resulting sim run shows a shocking development when run over the last 10 years. The chart shows a huge -49% drawdown in 2023.

When I investigated this, I was surprised to find the culprit was a $52-million market cap Chinese stock (CLEU) was somehow included in the 'revised' Mag 7 universe in July and August 2023. That stock contributed a -80% drawdown to the portfolio.

How can this happen? How did a $52 million chinese stock get into a top-7 market cap Nasdaq 100 stock universe??

portid=1845181

Thanks,
Chawasri

9

You're using the NASDAQ Exchange universe, not the NASDAQ 100. That universe includes plenty of ADRs. For a brief time in 2023, the market cap of CLEU, according to our data providers, exceeded the market cap of most other NASDAQ Exchange stocks: the price went up to $46,000 per share. It was never in the NASDAQ 100, though, because it's an ADR. And ADR market caps sometimes go awry because of data discrepancies.

10

Thank you much, Yuval! You are correct. I find that as I get older, details seem to fall through the cracks occasionally. I'll have to be more diligent in the future.

Chawasri