Forecasters of the stock market

Manage
41

Tom, Thank you very much for your post

Very interesting link

What to buy and when to buy also can look like a kind of timing, and we all do that for sure

What Danny and other are doing is not classic timing according to price, but he also look at the fundamentals and earning so yes, sure it can work

People will always use even funny timing systems like buying and selling according to months for example
People will also buy lotto ticket and some even will win, even with a very small chance it can happened,but is it a good investment?

I will never use something that worked in the past if I not understand why it works.

When you have a sp500 model and you timing the market (the sp500) with earning or other fundamentals it even make sense.

My points are

Is it smart to time small cap with large caps? Maybe, I am not sure

Is it smart to use something that worked in the past but you not understand why it was working?

And most important, if anyone can successfully and consistently time the market, why choosing individual stocks?
With all the flows Tom just mention, it make no sense to me

Isn’t it Better just to buy ultra-long and ultra-short on the market?

Amiran

42

Amiran,

Here is just one example of a pure SP500 timing system:
http://www.managedfutures.com/program_performance.aspx?fundtype=&productId=32812

On a stand alone basis it’s returns are so-so (albeit better than the market over 10+ years). And it’s had a somewhat rough 5 years. The reason institutions and pro. investors use these systems is they have zero (and often negative in times of peak market stress) correlations to the market. That is very helpful in total portfolio construction.

I am not suggesting that P123 users shouldn’t buy any stocks. Or should rely on any one system or style. The list of potential issues in long-short market timing are as great (or greater) than those with stocks. But not thinking about hedging intelligently across the market cycle or dynamically varying risk exposure in some systematic way is, in my opinion, a mistake.

Best,
Tom

43

There is a theory that explains why timing generally works better using large caps for generating a timing signal for small stocks. Large caps tend to lead the economy and the market. If large caps are having changes in sales and earnings, then the small caps that sell to the large caps will soon experience the same. Since the large caps have more analysts watching them, the good or bad news is often reported sooner for the large caps. A related factor is that financial data services tend to process info about large caps before smaller caps. So if a large cap and a small cap report changes in earnings on the same day, the data service will type the large cap changes into their database first. And the smaller caps will get typed when staff time permits.

This theory is not perfect, however, since in periods of growing uncertainity money tends to flow out of smaller caps into “safer” large caps. So if a period of uncertainity comes at a market peak, the small caps may head south before the large caps.

Yes. If one really thinks a timer works consistently, it makes sense to use the timer to trade Ultra ETFs. But if one REALLY thinks a timer works consistently why give up so much gain by uisng Ultra ETFs (which then use the futures market). Why not directly trade the futures market where one can get 5x, 10x or 20x leverage. With such a timer, in a couple years one should be able to buy a private jet to fly between multiple vacation homes.

I use market timing, but I do so to reduce drawdowns, not to improve gain. My view is that a robust market timer will reduce one’s overall gains by a bit (while reducing one’s drawdowns by a lot). If it were possible to find a timer that actually increased gains, such a timer would have already been found by guys and gals in hedge funds with muli-million dollar research budgets.

Regards,
Brian

PS: Just to clarify, I think a robust market timer will be able to reduce many BUT NOT ALL future draw downs. No timer I know of can be expected to predict a draw down caused by a terrorist attack.

44

Tom,
Thank you for the interesting link
Rough 5 years is understatement, this model did…well it do nothing in the past 5 years, when the sp500 go up about 100% at the same time!
(And they do ask for more then 800$ a month and 20% incentive fee)
They did however very well in the 2008 crisis .probably because they hold some shot on the market, but they continue to hold the short when the market tread has change, and I think it a good example for bad market timing

Brian
Your post explain partly why it is hard to make a good timing
The first example explain partly the different in the performance timing between large and small cap stocks
And this different are when the news or economics are the same for both of them

And the second parts of your post make my point even stronger,
Sure why not use the futures for short and long, and buy all the word

But, if you use the timing just to reduce your dropdowns
You can invest only part of your money
Or use some hedge all the time, this way you be sure, it will protect you when some unexpected event will happen

All of you make interesting argument for market timing,

Maybe some sectors tend to predict the market direction
I will look into it more and maybe come with a system of my own :wink:

Thank you all very much

Amiran

45

Amiran,

The model I showed you (which I don’t invest in, I build my own timers), has maintained zero correlation to the SP500 and high negative correlation in times of market stress. It’s one option people can look at as a hedge.

We look at the world very differently. That’s okay. And, there are many ways to hedge. I try to use a lot of them.

You don’t like ‘market timing’. And I wouldn’t be that comfortable with a constant TLT exposure as a hedge. But, this could work. TLT has had high upmovements in periods of recent market stress. And, of late, has maintained negative SPY correlation.

But, in the past 5 years, TLT has returned -15.89%. We would have also paid some margin costs, maybe 10-15% more. So, we would have lost 25-30% holding TLT over this period. The above timer outperformed that by a wide margin.

This brings up the fact that many / most hedges (long volatility strat’s) cost money in major bull markets where vol is low and/or falling. That’s hedging life.

And, I hedge some with TLT as well. But, for the most part, I try and ‘time’ the duration of bonds I am in at a given point (from SHY, to IEF to TLT) based on yield curve structure, total yield and recent price movements.

So…to me, raw TLT exposure is potentially a very risky hedge in the current environment.

Some of the factors include:

  1. The correlation between TLT and SPY varies over time, and sometimes is highly positive for certain cycles.
  2. And long-dated bonds are very sensitive to any news around interest rate rises - and rates are much more likely to rise than fall in the coming 5 years.

So…in most future years (barring major crisis and flight to US$), the going forward ‘upside’ is 2-3% (wiped out by margin costs)…and the down-side is potentially 20-30% for every real (or hinted at) 1% rise in interest rates.

So…in my mind a constant TLT hedge is more risky than shorting the SP500 (or an appropriate, strategy matched style index) when the index in a strong downtrend. But, I am holding TLT now. With a timer. And, it is a valid strategy and I accept it.

I wish you luck.

P.S. In 5 years, if P123 doesn’t let people delete records of systems from it’s PIT, R2G database, we’ll start to have some real data on mkt. timing and hedging’s impact on P123 strat’s.

Best,
Tom

46

By the way, I do agree that since 1980’s, US bonds have been very good diversifier in peak market stress times.

See (for example):
http://www.vanguard.com/pdf/s130.pdf

P.S. I do agree with you that the only way to be ‘sure’ of not losing money is to either a) leave some money in cash or b) always hold a hedge that is guaranteed to be inversely related to your long portfolio. It’s not clear that TLT or GLD will always maintain that negative correlation in times of severe market stress. They have of late. So have managed futures (most).

But, an SP500 timer definitely may not maintain that relationship. It could be long when the ‘bad news’ hits. However, to me, I use and trade timers. I’m okay with the risks. And I hold cash. And trade TLT.

47

Tom,

Different opinions and views are what make this forum so interesting
And I thank you for it

Sorry I thought the model at your link, was example of market timing to gain better then the sp500 performance and not as a better hedge to simply use TLT-long term bonds

Long term bonds and stock influence allot form the interest rate, but in inverse way of each other, that why I use it for hedge

If you look at yahoo and compare the TLT to the sp500 in the past 5 years

You will see almost perfect inverse correlation

And you will also see that the TLT did +7% in the last 5 years and not -15.89%
(Margin costs you can add to both systems if you use it,)

Most of us play for the long run, and like you said,
I guess we need to wait and see, if or what system or timing works best

I wish you luck

Amiran

48

Amiran,

I started my model at wrong date. But, P123 shows 4.27% annualized TLT return for trailing 5 years:
https://www.portfolio123.com/port_summary.jsp?portid=1174691

Correlation is a minus .5%. Not bad as a hedge in this period. But not my highest weighted choice going forward. I do think TLT is a good short-term play right now.

P.S. I do know someone who day trades the SP500 and Nasdaq futures, and turned $100k into over $3MM from 2006 to 2011. But, he’s not public or professional. He just trades his IRA. It’s strictly and intramarket system. No overnight positions. But, he’s had giant DD’s (over 60%). Just not when the market did.

Best,
Tom

49

On a long horizon everybody times the market since we’re born and die :wink:

But kidding aside, there are two ways of thinking about these issues that I find helpful.

First, there are 3 states of confidence on market timing. 1) Statistically confident it works 2) Statistically confident it doesn’t work and 3) Statistically unsure. Because we only have 2 big recessions in the 1999-current dataset I feel that the MA(21) analyst and other timing tools fall under 3), which still means they MIGHT work going forward.

So what is the downside cost if they don’t work? If we were perfect traders (perfectly buying/shorting daily lows/highs) then the timing would reduce our profits a bit more than just the lost time (market takes the escalator up and elevator down). If a system is profitable without market timing, adding timing can only subtract which is a good gamble for increasing risk adjusted returns. If the system depends on market timing to be profitable (i.e. cannot at least match the b&h benchmark) I worry the theory of the system is on uncertain foundations and am not interested in risking real money.

Second, shorting stocks is HARD so I find trying to make money by timed shorting is also hard. Companies have active managers waking up and trying to grow sales, profits etc unlike commodities which are pure supply and demand (with supply changes taking years) so trading long/short commodity futures I find MUCH more symmetric than stocks. Since bonds tend to go up during bear markets, bonds are more forgiving of timing errors than shorting so parking money in bonds should be more robust to timing errors (I’m not talking about being market neutral by maintaining a constant benchmark short).

This is the thinking I’ve used to (hopefully) add forward robustness to timing in some of my sims.