mindmoneymarkets

Days since 200 day highs

I found a recent post by Woodshedder on the behaviour of markets after 200 day highs and 200 day low absolutely fascinating. Well worth a read. 

I was interested in this conclusion: 

"The market making new 200 day highs is carried by its own momentum. The momentum allows the market to shrug off bad data and investor fear and uncertainty."

It got me thinking...could this provide some sort of long term trend filter?

It seems so.

Is it just the time of year?

As European politicans cut short their grande vancances to attend to the latest crisis, it's worth having another look how the Halloween effect is once again playing out. 

The theory is that markets perform best between November and May and this has certainly been the case going back to 1990 on the S&P 500. 

The chart below shows three investment strategies: 

1. Buy and hold the S&P 500.
2. Buy in the close in November, sell the close in May. Cash inbetweem return of 3% PA average. 
3. Buy the close in June sell in October. Cash inbetween return of 3% PA. 

Nov-May does give you more months in the market, but they do seem to be the right months. Food for thought...

What the crash says about the UK pension apartheid

For me, the recent market shakeup has highlighted the growing pension apartheid in this country. There is a clear division between…

1)      The pension haves (final salary pensions usually public sector) and

2)      The have nots (everyone else on defined benefit/ personal pension).

I won’t go into the politics of it all here, but following the recent turmoil, George Osbourne might find he is in a stronger negotiating position going into the next round of public sector pension negotiations.

As companies and governments close off their final salary schemes, there will be more people shunted onto defined contribution type plans which depend not on the number of years of your service/ final salary, but on the performance of your investments in the long run.

There are many problems with this:

Value + Momentum Update (PE10)

After reading this post from dshort musing that stock markets are looking a bit rich, I thought I'd update on the performance of the PE10 + momentum.

You can view my last update here and my original post here. 

The basic idea with this model is to use Robert Schiller's Cyclically Adjusted rolling 10 period PE ratio as the value measure. I've used a reading above 20 as the point of being overvalued. Below you can see that historically the S&P 500 has actually spent more time in with a PE 10 reading below 20. 

The second part of the model looks beyond just looking at value. The idea being that markets can be expensive and carry on going higher for years. The best example is the UK Housing market which was expensive for years but kept on going up and up. Even as early as 2003 people were saying they had sold their house ahead of the coming crash, but that was too soon. Prices have dropped, but are still above 2003 levels, so selling based purely on something being 'expensive' is not enough. They might get back down there mind you, eventually in an inflation adjusted basis.

So you need a momentum component. For this I used the 10 month moving average of the market as suggested by Mebane Faber. 

The rule is to buy if the market is above the 10 month MA and sell if the market is below it. The kicker is that this rule only kicks in at points when the market is over valued (PE10 over 20), otherwise you just buy and hold. 

So after all that waffle, here's the updated performance for the longer and shorter term: 

Right now the momentum is almost relentlessly pushing higher, while the valuation level makes it look expensive. The Value + momentum model would keep you in the market until there is a pull back. 

 

Always read the small print...

Hi and Merry Christmas & Happy New year to one and all.

Thanks for reading this blog over the past year. I’ve enjoyed sharing my research over the last year and I’m pleased to be back in the hot seat now my move back to the UK is finally bedded in.

The New Year is a time for reflection on the year’s successes and mistakes. I feel I learn best from my mistakes, especially when I readily admit to them rather than trying to self justify.

I won’t bore you with the details of my first mistake. This was to chose Virgin Media as my internet provider. It took 8 weeks and multiple calls to their call centre to finally get set up only to find the connection not even at dial up speed. I’m there now, but if your work depends on an internet connection being set up promptly then I would seriously think twice before signing up with Virgin Media.

Ok on to my main mistake of the year (actually from past years, but my fault for not checking it sooner).

Always read the small print with investments.

A few years ago I belatedly set up my personal pension. This was before I’d done so much serious in depth research into the impact of trading & investment costs. I wish I knew then what I know now.

I knew enough to keep costs low. I chose a well balanced portfolio of funds with the average annual management fee being 1%. This is actually relatively high compared to other options available, but they were certainly the lowest fees on offer amongst the various funds presented to be by the Independent Financial Advisor. I could have easily gone for funds with annual fees of 2% or higher. I also made sure that there were no initial investment fees in the funds themselves. What I didn’t check was the small print regarding management fees.

Following this excellent article in the Telegraph on high fees UK investors are paying compared to other countries such as The Netherlands, I decided to take another look at my pension.

Here’s a choice quote:

“If a typical Dutch and a typical British person save the same amount for their pension, the Dutch person can expect a 50pc higher income in retirement.”

Why is that? Are the Dutch and the Danes better investors? No, they simply pay lower fees, which have less of a negative compounding effect on your investments.

The original study was commissioned by the RSA with warning comments about the cost of ignorance on costs such as: “When people hear that they are being charged 1 per cent or 3 per cent, they think they are being charged 3p in the pound or 1p in the pound, and think 'that's fine'. But because of compounding, the pension costs actually add up to 40 per cent"

Warning bells

I had a vague memory about exit penalties from the pension if a decide to come out of it early or switch. However I had never paid much attention to how this was actually structured.

My mouth nearly hit the floor when I clarified this recently.

The pension charges me 5% PA on top of the Annual Management fees for the first few years then it reverts to no extra fee and rebates me by 0.4% for the remainder of the pension.

In other words, for the first few years of my pension I have to pay 6% PA in fees for the privilege of getting fees of 0.6% for the rest of my time.

Although they are in the documentation, these fees were certainly not explained to me in these terms by the advisor. Put another way, it would take me 21 years for the benefits of the 0.4% rebate to start beating a set 1% every year.

The exit penalty is a fixed £362 to boot and I have three years of paying 6% before the rebate starts to kick in. I have three years of paying 6% left.

What to do?

There are various pension fund providers in the UK that provide rebates such as http://www.cavendishonline.co.uk/  allowing you to pick up fees in region of 0.6% straight off the bat. The January edition of Which? Magazine also had a great table showing the various providers and the discounts they provide.

The other option is a SIPP provided the set up costs are low. http://www.sippdeal.co.uk/ look a likely option. The disadvantage with a SIPP is that it is designed around individual purchases not regular investments so I would not be able to make regular contributions each month as the transactions fees would be prohibitive. Instead I would have to make annual purchases and pay fixed dealing costs.

The advantage is that a SIPP opens up low cost ETFs and Vanguard funds we fees in the region of 0.2 to 0.4%.

It immediately became apparent that the decision to switch out is a no brainer.

Say I have £10,000 in my pension pot. 6% of that is £600 lost in fees to my current provider every year for the next three years.

If I switch to another provider and pay 0.6% for that year, then it would be £420 lost in fees and charges, £180 better than keeping with the current provider. I’m up in year one with money back to me over the following two years.

However, if I can find a suitable SIPP provider, I may be able to push annual fees down to 0.3% and by making annual purchases I can keep transaction costs to a minimum.

This could make a dramatic difference over the life of my pension.

I would be £7,000 better off on a £10,000 pension switching to a provider giving me fees of 0.6% from day 1 compared to my existing provider. This assumes 5% growth P.A. and contributions rise with inflation. However, the advantage of paying 0.3% per year and a fixed £60 in fees is huge.  Switching to such an arrangement would be worth an extra £40,000 over my current provider.

Aside from the potential for extremely low fees through a SIPP, I like the possibility of accessing ETFs and trackers which I expect to have low hidden fees.

Hidden fees reduce costs even further.

Terry Smith explains more

“There is a justified focus on the Total Expense Ratio (“TER”) of funds which include those expenses which the manager charges to the fund rather than simply on the fund management fee. However, there is one major cost that is not charged to the funds-the cost of dealing in the underlying investments.

This is not insignificant given that, according to FSA research, the average fund manager in the UK turns over their fund 80% per annum. This adds three layers of additional costs: 1. The commissions charged by brokers and investment banks for dealing; 2. The difference between the bid-offer spread for securities sold and bought; and 3. The fact that no fund manager has sufficient good investment ideas to warrant buying and selling 80% of your investment portfolio per annum.”

These costs are generally felt via under performance rather than management fees. I’ve been checking a few ETFS and they seem to underperform the benchmark by a margin that is in proportion to the management fees as expected, but I shall be triple checking this.

Don’t be fooled like I was.

I hope my sharing of this mistake is of use to you. I know the pension systems are different across the world, but from my experience, it does seem that the UK system is highly skewed against investors.

This will become ever more apparent in the coming years as more companies shut down final salary schemes and even the public sector may follow suit if desperate. From 2012 financial advisors will be banned from taking commission on investments which I think will alleviate some of the problems, but not all of them. They will still receive ‘trail’ commission from investments made before 2013.

Of course, this is not financial advice. I hope by sharing my frustrations, it sparks your own investigations.

More musings to come…

Further Reading:

http://www.thersa.org/about-us/media/press-releases/going-dutch-how-to-double-the-value-of-british-pensions

http://www.independent.co.uk/money/pensions/revealed-the-scandal-of-how-pension-providers-rake-in-the-money-2157940.html

 

 

Reading & Requests

Just feeling my way back into the markets after a nice road trip from Vancouver to Calgary and back through beautiful British Columbia. Staying a night in a teepee in Golden was certainly a highlight.

A few interesting articles out there to share:

• I've been wondering about hedging vs stop losses for various strategies for some time, especially after Rob Hannah touched on the issue in his subscriber newsletter. Do stops help or hinder trading strategies, especially those that use mean reversion? Of course its drilled into us that we need stops, but in the case of some methods, they can limit returns. In which case hedging your position or portfolio via options might be better. Condor Options has a timely article on the subject, using VIX based ETFs as part of the hedge. Another article on using SPY put options here. 
• Here's the request - Can anyone point me in the direction of a source for historical daily or even monthly option price data? I'm thinking the SPY or SPX initially one month out of money with the strike roughly 25% below? Rolling front contract. Perhaps 12 month might be easier to get around time decay.
• On the subject of hedging, I tried to recreate the Condor Options strategy, but had little success, though I did spot something else which I'll get to in a minute.
• Bad news for UK savers with the NS & I withdrawing Guaranteed Inflation Linked Bonds. These were the only guaranteed way to beat inflation on the market. Some good and not so good alternatives outlined here. Notable is "If none of these solutions sounds attractive, David Black, at the financial information group Defaqto, suggests looking at whether you can overpay your mortgage, assuming you have one." 
• Great series from Market Sci on the Golden Cross. 
• Gives me an idea to investigate a strategy that staggers entry based on following a number of different crossover methods. 
• More on crosses from Schaffer, this time the death cross (Which Market Sci also examines)
• On another note - Some challenging posts from Barry Ritholtz on climate change. I've read so much conflicting information on this I just do not know where the balance lies. After hiking around BC and Alberta and seeing all of the glaciers seriously receding, I'm confident things are getting warmer. Is this man made or natural cycle? I don't know, but in my opinion it's better to not take the chance that it is an unnatural cycle. Nature is full of stories demonstrating the law of unintended consequences. 
• And finally more nonsense on managed funds over passive funds from the Times. The writer argues that BP dropping so much is a case for managed funds. Utter drivel. The writer assumes that the average active fund manager was nimble enough to get out of BP before the bulk of the losses set in and perhaps more importantly, will get back in to BP at the right time. Those same managed funds were not so nimble during the credit crunch.
  

VXX Mean Reversion Hedge

I got an idea from the aforementioned Condor options hedging article to look at the hedging capabilities of the VXX (ETF based on VIX futures - i.e volatility). My initial hypothesis was that VXX could be used as a hedge on long trades that use MR.

To cut a long story short, I couldn't find an edge with it. Part of the problem is that the VXX has only been in existence since Feb 2009, a period that has been largely benign and bullish with the exception of the volatility seen since April. So any hedge will have had minimal impact and since April there isn't enough data to go off. 

The VXX has been in a steady decline as volatility died off following the credit crunch, potentially eroding the value of any hedges put in place.

Condor options use the vix futures as a proxy, but I don't have the data, can anyone help with this?

Hedging using VXX or VXZ may indeed have merit, I don't doubt the condor options study, it's just that I can't recreate it personally.

I did notice something else though.

Since 2009, returns for mean reversion have been below average. Largely because there's been a lot of upside trending and not too many downside spikes like we saw in 2000 and 2008.

The short trades have contributed very little to the overall total. Take the DVSC indicator for example from DVindicators. Since Feb 2009, using the settings of long <0.5 and short >0.5, long trades have returned 41% and short trades just 3%.

However, the VXX still has a role to play here. Condor options mentioned that volatilty begets volatility, so I wondered if this could be used to modulate the performance of the DVSC.

I applied a 20 period percent rank (PR) on the VXX and examined the reaction of short and long trades. I was most interested in extreme readings (0.8 and 0.2). I found that extreme readings in both directions were pretty poor for both long and short trades, with the bulk of profits coming when the VIX was inbetween a percent rank of 0.8 and 0.2 (trending but not that strongly)

At first glance this was not what I expected, I expected a high vix percent rank to lead to better long trades and vice versa. But I realised I was forgetting the sort of system being traded which is in itself mean reverting and in some ways a volatility hedge (At least in terms of giving you short side exposure).

I applied a filter so that the system only traded if the VXX 20 period Percent Rank was above 0.2 and below 0.8, otherwise return 0%. This had the biggest impact on short side trades so I experimented with a different rule (with the benefit of hindsight) that if the 20 VXX PR was below 0.2, don't just scratch the trade, reverse it and go long instead of short.

The below 0.2 rule makes sense because of the data set. The VXX has been trending lower for most of 2009 and Q1 2010. In this case a reading of 0.2 meant low expected vol which turned out to be the case, hurting short trades because a low VXX reading generally went hand in hand with higher markets.

The red line (Should read "Reversal VXX hedge" is for the DVSC strategy that goes long instead of short on a VXX percent rank of less than 0.2. The purple line is the binary method of not trading the DVSC if the VXX 20 period PR is below 0.2 or above 0.8.

Right now I don't know if this is something meaningful, or a factor of a very limited dataset going back to just 2009. The 'reverse shorts' rule may only work because the market rose more than it went up since then. Having said that, it has helped improve the basic DVSC indicator since April.

I have tested the same filter on the DVB, DVAR and RRSI and go similar results, but I think I'd need more complete VIX futures data to be able to test this fully going back to 2000. Food for thought though.

Active vs passive funds

If you haven't seen it, there's a nice argument building on FTfm (Financial Times Fund Management) about the merits of Active vs Passive investing.

I was particularly intrigued given the new UK coalition governments proposals to increase capital gains taxes on shares to 40% and perhaps 50%. The is a backlash against this, so the plans may be watered down when they come out. Any increase in fees will no doubt push more money into tax efficient vehicles such as ISAs or pensions so the choice of what to do within those wrappers becomes ever more important.

Short term/ active trading is one thing -  What I'm talking about here is the supposedly safer, more longer term means by which we save for our future/ retirement. I've covered a few points on this here.

The argument over active vs passive investing will continue for years I'm sure. Naturally there's an element of both sides talking their own book, but I come down on the side of team passive investment. 

Trackers are more volatile .....

One thing that gets my goat is misinformation. I can't find the original link but I read an article in the Telegraph proclaiming that many financial advisers would be sticking to actively managed funds because they were less volatile. This article from the FT came with this quote which is just unreal: "

“One could argue that [passive] exchange traded funds are only suitable for investors who can sustain a 50 per cent loss in a year, given the fact that in the last decade the S&P 500 has suffered two severe drops, one of 46 per cent and the other of 56 per cent.

“One of the most basic things people expect from an investment adviser is that when the market heads south, someone will step in to protect them and help them to recover. ETFs, however, can never meet this expectation.”

Nonsense, he's comparing apples with bananas

A) Who in their right mind is exposed 100% to a single market for their long term investments? There's tons of free ideas out there on how the lay person (I firmly include myself in this) can diversify their long term investments such as investing 60% in stocks and 40% in bonds. The strategy historically performed as well as the stock market alone but with dramatically reduced volatility. With the use of index trackers you could do this with a cost of 0.5% a year or lower compared to a 5% load fee and 1%+ annual management fees from most managed funds.

B) The comment implies investment advisors are superior market timers....Really?

Here's some choice findings that dispute this from the same FT article.

"Yet the figures do not seem to bear this out. The monthly annualised standard deviation, a measure of volatility, of the UK’s FTSE All Share Index was 19.2 per cent during the three years to April 2010.

Yet the average actively managed fund in the UK’s All Companies sector had a standard deviation of 20.9 per cent, ie was more volatile, according to calculations based on Morningstar data."

Alan Miller, partner at SCM Private, a UK wealth manager that constructs portfolios purely of passive ETFs, labels the claims that passive funds are higher risk as “utter nonsense”.

“You are always going to have a more concentrated portfolio in an active fund, and with a more concentrated portfolio you would expect to have higher volatility,” he says.

“When one combines this with the fact that the average UK active fund typically underperforms the market by 1.3 per cent per annum [the average during the five years to February 2010], one can see that most active funds combine more risk with lower return.”

Nowadays individual investors can play the active manager with their own timing models. There is a fair body of evidence to support the idea that momentum models can at least provide superior risk adjusted returns.

Heck these may not play out as they have done in the past, but I'd rather make those mistakes myself as I'm doing with my own rotation portfolio than pay someone 5% upfront for the privilege of making the same mistakes.

The other side of the coin.

In defense of active management, I'm sure there are some genuinely hard working, talented professionals out perform thanks to skill. Something that doesn't make sense is the body of research on the success of simple diversification portfolios and rotation models vs the under performance of actively managed funds.

Do any funds apply these simple diversification or rotation models, or are they deemed to simple? Perhaps they are used, but only work with the benefit of hindsight? Perhaps so, but then we get back to the issue of risk. Where do the risks lie with passive/ index investing vs active management investing?

For me it comes down to this - Irrespective of whether you buy and hold with diversification or rebalance based on momentum, for the average investor you have a choice. With passive funds you're guaranteed to not out perform the benchmark, but in compensation you'll pay lower annual fees. With active funds, there is no guarantee you'll even track the benchmark let alone beat it. However you are guaranteed to pay more each year in fees which can eat up to 25-30% of your gains in the long run.

As a risk/ reward pay off in these terms, passive investing makes sense to me.

Further reading:

• Fund charges exposed as fees outstrip returns. 
• Tracker funds take a tumble. 
• Don't tar all actives with the same brush
• Data subverts attack on index trackers.
• Twisting the facts on active management.
• Active managers - Skillful, lucky or useless?

Three FTSE 100s

As a follow up to my previous post, I thought it would be interesting to share the following:

Here are the trading results from three different FTSE 100 vehicles all trading the same method - The DVB buying below 0.5 and selling above 0.5. 

FTSE = FTSE 100 cash index
ISF = UK listed ETF of FTSE 100
EWU is the US listed ETF of the FTSE 100

Software used: DV Indicators

Naturally the EWU is a slightly different beast as its listed in $ not £, but even so, its striking that these results are superior to the other vehicles.

Could it be that the EWU trading US hours is more closely aligned with the more tradeable SPY (mean reversion type pattern friendly), thus increasing returns on a FTSE 100 vehicle?

Update:

I managed to get hold of some FTSE 100 futures data which shows the futures trading in line with other UK based FTSE vehicles. FTSE futs are listed in £ but trade later with US markets. This appears to be the only obvious remaining differential. Perhaps it is the £/$ making the difference after all? Next stop.... attempting to adjust returns for £/$ movements.

Thoughts on the FTSE

Just thinking out loud with this and would welcome any other opinions.

I've been doing some extensive testing of David Varadi's new product and am impressed (to say the least), although have only scratched the surface so far. One thing that puzzles me me is that the FTSE does not respond to all the strategies/ indicators in the same way as the S&P 500. Naturally, you wouldn't expect it to. All markets have their own dynamics and this in some ways poses a wider question - how do you assess the dynamics of a particular market and how its characteristics differ from other markets?

In other words, how do you best optimize a system or indicator to a particular vehicle without curve fitting the hell out of it?

I don't think its as simple as saying 'The FTSE doesn't like Mean Reversion' etc, I think the differences are more subtle.

Some thoughts:

• The indicators work reasonably well on other US markets & ETFs, so I don't think everything's been optimsed to the SPY, although this certainly seems to the king in terms of repeatable patterns.
• I think its more to do with the nature of the way the FTSE trades itself.
• Two hypothesis I'd like to test in the next few weeks:
• • Is the FTSE impacted by the time delay of US markets remaining open after it closes with the strong SPY patterns overriding anything from the FTSE? I might be able to test this by looking at the US FTSE ETF, or using SPY signals to trade the FTSE.
  • Does the pound have a role to play, particularly the GBP/ USD. Sure most US companies are international, but I don't know if the % of their income from overseas is as large as the FTSE 100 constituents. In other words, the GBP/ USD could be diluting the impact of any core repeatable patterns (mean reversion etc). I wonder if adjusting the FTSE to account for GBP/ USD movements would impact on returns from Mean reversion type systems.

Could just be that the FTSE is a bugger to trade. In the grand scheme it shouldn't matter too much and I know theoretically that it's better to focus on what does work (the SPY etc), but I hate leaving a puzzle like this unanswered.

NB Corey from the DV indicators team has sent over a fascinating correlation of returns chart for the FTSE which does provide some clues. I'll see if I can have permission to publish it here.

Short term indicator rotation model

Following on from this post on the best Short term indicators, I wondered if would be possible to create a rotation model on indicators just like you would momentum stocks. I.e. you back the to performing indicators only each month/ week/ year and rotate your strategies accordingly.

Well from my small sample size of three systems (all of which are very similar it must be pointed out), it turns out that this idea might have potential.

I did a sum total for each indicator (DV2a, Dv2b, Ultimate Oscilator and the market itself, SPY) for each month.

I then took a rolling total of 5 month returns from each indicator + the market. Follow only the indicator with the highest total over the last 5 months for the next month. Repeat process at the end of every month. If the market is the top performing activity then stick with that.

All indicators set to >0.8 as the sell trigger and <0.2 as the buy trigger. No commissions/ spreads - I'm looking for a proof of concept at the moment not a trading system yet.

Here's the rolling total monthly returns. The 5M Av is the rotation model (the 2nd best line if you can't see the colours).

This doesn't tell the full story though. I attempted to show the risk adjusted returns by comparing the standard deviation to the total returns. I then divided the returns by the standard deviation. SD here is the raw monthly standard deviation, not a % of the SPY's value.

	DV2(a)	SPY	DV2(b)	UO	Rotation
StDev	4.213346	5.745971	5.839076	4.70817	5.988967
Return	160.59	36.44	282.98	218.78	274.25
SD adj Return	38.11	6.34	48.46	46.47	45.79

Again this doesn't quite tell the whole story though because the Stdev measure used can penalise a system for upside volatility. So let's have a look at downside standard deviation only. 

Negative SD only	DV2(a)	SPY	DV2(b)	UO	Rotation
StDev	2.570973	3.975319	2.723411	2.529702	2.323051
Return	160.59	36.44	282.98	218.78	274.25
SD adj Return	62.46	9.17	103.91	86.48	118.06

Now we can see the real potential with the model, it reduces downside volatility while keeping returns steady.

Next I wondered what would happen if you hedged your bets by backing the top 2 performing systems next month. This might mean investing fully in the SPY for one half and trading Dv2 for the other half of your funds.

Interestingly, returns are again below the top performing indicator, but this time have a much lower Standard deviation even before we isolate downside volatility.

	DV2(a)	SPY	DV2(b)	UO	Rotation
StDev	4.213346	5.745971	5.839076	4.70817	4.479649
Return	160.59	36.44	282.98	218.78	252.73
SD adj Return	38.11	6.34	48.46	46.47	56.42

Here's the hedged version with downside volatility isolated. 

Negative SD only	DV2(a)	SPY	DV2(b)	UO	Rotation
StDev	2.570973	3.975319	2.723411	2.529702	1.822053
Return	160.59	36.44	282.98	218.78	252.73
SD adj Return	62.46	9.17	103.91	86.48	138.70

Here we can see that the monthly standard deviation drops markedly while keeping returns pretty constant. The risk adjusted return therefore really stands out.

So encouraging signs that this idea has legs. For the record the most common activity used was the SPY itself. 

No doubt there is a more sophisticated relative strength/ momentum model than the one I have used that could be worked out.
Would the model still work with other indicators? (non mean reversion for example?).
What if you throw in some indicators that have since stopped working (is there a survivor ship bias here?)

Lots of questions, but at least it seems like this could be a fruitful line of research.

Perhaps the most encouraging thing I've seen is this graph from the FTSE 100. Same settings as SPY. The short term indicators just don't seem to work as well on the FTSE 100, perhaps because of the delay in reacting to US markets. However, the rotation model still managed to pick out the one indicator that was working and rotate out of that when the others/ the FTSE started to improve.