Archive for February, 2009

Looking for the reversal

Image by Julián Contreras via Flickr

Just a little update here: VIX looks stretched here – and I’ve got a buy on a system based on the McCellan Oscillator which, in the past, has had a success rate of around 70% but started to fall apart in October.  So, bottom line, I’m looking for a positive day tomorrow.

Read Full Post »

Mutual fund
Image via Wikipedia

Michael over at Marketsci asked a great question in his latest post – why do mutual funds continue to show no ability to think outside-the-box despite the abundance of alternative investment approaches that are available publicly (either from individuals or via public research) or from individuals that they could potentially hire to help run their fund (particularly in the quant area).  I thought I’d take a post and attempt to answer Michael’s question.

But first, a brief story.  In 2003, I was working in New York City at a start-up and just beginning my “quant journey” when our company was sold.  As Chief Technical Officer of the company, I was out of a job as there was someone at the new company that already held my position.  I was given a nice consulting contract and time to figure out what I was going to do next.  I was approached by one of the original investors my startup who was starting a hedge fund.  He felt my background in technology would make me a good analyst on technology-related stocks (I’d also told him I thought Apple was a buy in 2002 and he’d made a lot of money off that call).  When I told him I had no idea how to be an analyst, he offered to teach me.  Thus began what I can only describe as my on-the-job MBA training.  We were a fundamental-oriented shop, and I learn everything about fundamental analysis.  We did ok – not bad, not great.  And then my boss was offered a job at another fund, and he decided to shut down our fund and get started over at the other firm.  He’s done very well and is still one of the best investors I’ve ever met.

The question then, for me, was what to do next.  I wanted to get into another fund, but my background and experience was unorthodox to say the least.  A few years later I got an introduction, via a family friend, to a large mutual fund company that was looking for a small-cap analyst to join their team.  I was somewhat excited to go in – it would be a job in the right sector – but somewhat concerned about what it would be like working for a mutual fund.

I studied up on the fund for about two weeks.   The fund seemed to have changed strategies multiple times, and had now settled into a strategy of equal weighting sectors, and then using stock picking within each sector based on both quantitative measures and analyst recommendations.  It held 100 stocks across all the sectors.  All this sounds somewhat interesting, but the performance of the fund, to date, had been terrible.  An investor could have outperformed the fund in all time periods simply by buying and holding IWM.  Adding insult to injury, the fund had a huge sales fee of 2.75% (that’s an upfront cost to coming into the fund) and a very high expense fee for the fund itself.  So not only was the fund underperforming a simple index strategy, it was doing at a huge cost.

The interview itself was cordial but I knew, going in, that I was a long, long shot by any stretch of the imagination.  The fund manager asked me his first question: “how would choose stocks for this fund?”  I had given this question a lot of thought, and even run some simple tests to understand what the portfolio would look like.  I proposed to do a momentum ranking of sectors and then choose to overweight the specific sectors that were outperforming over a 12-month time period, and then select stocks within those sectors based on a variety of different criteria, including some fundamental-based quantitative measures as well as some price-based measures.  I then showed what the returns could potentially look like.

The fund manager didn’t seem to really care.  Perhaps it was because he felt the interview was pointless in the first place (because my background didn’t match with his expectation), or that he just so disagreed with my approach, but he then went on to tell me why my approach wasn’t appropriate and why he thought that their approach was superior.  Now, I chose not to be too confrontational with him on this because it was, after all, a family friend that had introduced me, and I didn’t think there was much to be gained by defending my approach too much.

But what I wanted to say to him was this: let me get this straight – you’ve underperformed the index since the inception of your fund while overcharging your customers for the privilege, and you’re telling me that my approach lacks merit? Instead I pointed out some issues with his current approach and refocused on the benefits of my approach – including citing research papers about the momentum effect.  Whatever – the interview was clearly over at that point as I could see him getting ready to move on in his day.

I asked him then what type of candidates they were looking for for the position.  He said that they were looking for someone with 3-4 years of buy-side experience or someone with 5-6 years on the sell-side (an MBA was pretty much implied here).  This was understandable at the time – getting someone from the buy-side to move over to this position in 2006, the height of the exodus from mutual funds to hedge funds, would be challenging – so he was mainly focused on the sell-side analyst community.

He ended by thanking me for coming in, and then suggested that my approaches seemed more appropriate for a hedge fund and that I should focus my job search in that area (remember that comment – I’m going to come back to that in a bit).

Ok – the story was supposed to be shorter and got longer – apologies – but I think this story demonstrates a lot of the issues with mutual funds today.  Let’s break them down.

  1. Small Gene Pool:  The mutual fund community generally hires people out of business school – not people with a background in sciences or any other background (Bill Miller’s degree in philosophy notwithstanding).  As a result, they’ve got a fairly homogeneous group of people that were all trained the same way.  The “craziest” they tend to get is to hire someone from the sell-side – which is interesting because the sell-side is a pretty good contrary indicator to stock performance in my experience.
  2. Reliance on a single approach: The mutual fund industry relies on the marketing idea that stock picking works (whereas research shows it rarely does) and that their particular group of analysts are the best stock pickers year after year.  Watch mutual fund commercials – they all have the same tone: “Experience, driven by the long term investment view, top team, etc.”  They all usually have the same kind of images as well – a team of people rowing together, or some sort of relay race, etc.  Yet any form of analysis of the industry shows the opposite – that any person would be better off with just a collection of index funds with low expense ratios.
  3. Strategy Scaling Requirement:  Mutual funds (and hedge funds) have a large issue in terms of scale.  For a mutual fund to make money – because they make their money on fees – they need to get big.  That means they can’t have very active strategies.  And it limits the companies they can invest in.  And it means that when they buy or sell, they are the “dumb money” – they tend to sell at the bottom and buy at the top because it takes them time to get into and out of a position.  In addition, many of the quantitative strategies that people put forward on the net (including my own) simply won’t scale to the size of a $1b fund.  For the individual investor, this is the advantage they have over these big lumbering giants.  The same is true for hedge funds as well – a friend of mine has a strategy where he simply invests in small hedge funds when they have less than $500m under management and exits when they get bigger than that – because he knows that funds are hungrier when they are smaller and can be more nimble.
  4. Belief that quantitative strategies are more risky: Remember the comment in the interview “maybe your approach would be more appropriate for a hedge fund” –  very interesting.  What he was saying, in my opinion, was actually “your strategy sounds riskier, therefore my clients will be scared” (It could also reference the scale issue mentioned above or that he thought I was a risky individual in terms of my background).  What he failed to understand (or even ask me about) was the risk-adjusted return.  He was just focused on the idea – not the result – as was obvious in the way he was running his own fund.  He was in love with the idea regardless of the results.
  5. Lack of a burning platform: So this fund was able to charge a 2.75% sales fee in addition to the yearly fee, underperform the market and still managed to have some customers?  What that tells me is that the people they’re selling this shit to are either not engaged in the world of investments (pension funds, individual 401k participates), in the pocket of the mutual fund company via some relationship, or they are just stupid.  So what, exactly, would be their incentive to change the way the fund was run?  Exactly – there was no incentive.
  6. Fully invested: This one is easy – most mutual funds have to be fully invested – keep their cash at a minimum amount.  This is a huge disadvantage over the individual investor.  Now, historically they’ve made the argument that timing the market doesn’t work and buy and hold is the way to go.  In general they are correct that most people cannot time the market.  However, if you buy into the buy and hold argument, then you’re still better off going with index funds.
  7. 100 stocks or more = Indexing: Most mutual funds (including the one I was interviewing with) have to hold a rather large number of stocks in order to scale.  In my experience, I’ve learned that approximately 60-65% of returns come from the general market, 20-25% come from the sector, and the remaining returns come from the individual stocks.  But if you’ve got over 100 stocks in a portfolio, you’ll discover that the portfolio starts to become very similar an index (assuming you’re doing equal sector weighting like they were).  So they were, essentially, a closet indexer with a high cost.
  8. Skepticism of quantitative price analysis:  This is a generalization, but I’ve found that most of the people that run mutual funds think that analyzing price action or using some other form of quantitative analysis is either voodoo or stupid.  To me, this is why James Simons over at Renaissance can eat their lunch while charging huge fees.
  9. Belief that they can out-perform the market:  This is a basic one but I think it should be stated – if you completely believe that you can outperform the market through fundamental stock picking, year after year, then that is your world.  If they would do any research on the subject, it would probably cause a cognitive dissonance so large that they would disappear into a deep depression.  Or they would say “fuck it, I’m making money off these chumps”.
  10. Belief that they can understand what is going inside a company:  Enron is an obvious example of this, but I see it all the time – these guys actually believe that they can fully understand these companies from the outside.  Even after SarBox has taken away many of their tools.  To me, what they are actually doing is guessing what is going on inside companies.  And you can never really know as is evident watching companies both beat and miss on earnings.  I can remember sitting at my desk at the hedge fund, waiting for an earnings release and realizing that I was just hoping the results would come out as I thought they would.  Hope, as they say, is a four-letter word.

I’ve rambled on enough in this post – I’d love to hear from readers as to what are other reasons that I may have missed.

Reblog this post [with Zemanta]

Read Full Post »

The Skills Dat Pay Da Bills album cover
Image via Wikipedia

The Skills Index is back on sell again.  This is a tough market, no doubt about it.  Short term, though, I wouldn’t be adding to short positions here with RSI2 on the SPY down to around 3, and the percent of stocks in the SP-500 with RSI2 of less than 10 approaching 60% – pretty close to other times where we saw reversals.  So what to here?  Well, I’ll watch the trade during the day for a reversal.  Other than that, for short positions I’d look for a retracement on this move to establish positions, but honestly there really has been no strong trend (except the overall down trend), so I’d limit both long and short positions.  I think if you are trading here, keep it small.

Suffice to say that all the positions established in my prior post from MR1 and MR2 are underwater – I’ve got one new position coming from MR2 – QLD.


Reblog this post [with Zemanta]

Read Full Post »

winter reflection
Image by Qba from Poland via Flickr

Hey all – I’ve got some new setups from MR1 and my new, still in development MR2 that I wanted to share with folks.

MR1: EWJ, OIL – both buy on the open with a 50% allocation to each.

MR2: DDM, DIG, MVV, RSU, SAA, SSO, URE, UWM, UYG – you’ll need to look at each of these to determine your own position size – I use a custom position sizing on this.

Read Full Post »

Not really shocking based on the mini-rally we’ve been experiencing.  Anyone notice that SLX has been killing it?  We’re approaching resistance across the board – rising number of stocks with RSI2 of greater than 90 – so we are probably looking at a pullback next week.


Read Full Post »

Adapter between a female BNC connector and ban...
Image via Wikipedia

I don’t know about you, but I think Michael over at Marketsci has been publishing what I consider the best stuff on the web on adaptive systems.  It has taken me a while to figure out how to do this with my own tools (Mike does this all in Excel – which is amazing in itself), but I’ve put together a simple system for IWM (the Russell 2000 ETF) to play around with making a system adaptive.

In this case, the rules are rather simple:

Buy: Buy when the 9 day simple moving average crosses above a 6 day simple moving average.

Buy Price: Close of the day following a signal.

Sell: Sell when the 9 day simple moving average crosses below a 6 day simple moving average.

Sell Price: Close of the day following a signal.

Investment universe: IWM only.

Note that this system is trading rules that generally go against conventional wisdom – which is that a short term moving average crossing over a longer term moving average should be a buy, and the opposite for a sell.

Now, to make it adaptive, I’ve made it so that if a 250-day moving average of the equity curve of the above system is declining (value today is less than the value yesterday), I will adapt the system.

So how does it adapt?  In this case, again, to keep it simple for this example, I’ve made it so that it will simply trade the rule opposite.

Now, the equity curve and system here are nothing special – I’m just using it to get comfortable with creating an adaptive system.  It is pretty clear that creating such a system has more challenges than normal system development.  Questions to ask yourself are (and Michael, again, is the authority here):

  • What causes the adaption? Equity curve?  Consecutive winners or losers?  Daily returns?
  • How will the system change? Reverse the rules?  Go with a completely different system?  Modify the existing rules?

Below are the equity curves for the system – the top of the chart is simple IWM with the moving averages thrown on it.  The black line is buy and hold on the IWM, the blue line is the system as it trades using only the default rules.  The green line is the adaptive system.  The moving average on the blue line is the 250-day SMA that is used for the adaptation.  So, when the line is green, we’re in the default system, when we’re in the red, we trade the rules opposite.

This is really just a baby step on my part in this direction – and I have to thank Michael at Marketsci for both his contributions to the community around this core idea and for his personal help to me in discussions on the topic.


Reblog this post [with Zemanta]

Read Full Post »

The logo for the first American Survivor seaso...
Image via Wikipedia

Market Monk has been doing an interesting series on Survivorship Bias – and how much of a problem it can be.  What shocks me is that not one developer of system testing software has dealt with this.  Here’s the deal – if you’re testing on the Nasdaq 100, your test is likely to be invalid because of the stocks that have gone away.  Or if you’re creating an index based on the S&P 500, then that index will likely not be valid going back in time, say, a year before the index has problems.

So what’s the answer here?  One possible answer – create a custom index (with all components) that includes all stocks that have traded in the index.  Now, you can get a list and data for all the delisted stocks – but the issue is creating an index that either includes all of those delisted stocks, or have the list be dynamic – meaning that it updates as changes are made.  As of this writing, I don’t think one data vendor or system software creator has a tool to deal with the issue.  I’ve been chatting with the folks at Norgate Premimum Data services – it sounds like they might be working on something to deal with the issue later this year.

Reblog this post [with Zemanta]

Read Full Post »

Older Posts »