Friday, March 30, 2007

Trends in Losses and Wins

Crazy day... Can't tell you all the craziness or I'd blow my cover as it's out there in the news media :) On top of that, Snork Maiden got an interview with a university in Europe for a post-doc type of position. Pays better than any US post-doc but prices and taxes are high there. I reckon 50/50 she'll get it. Maybe better than that as one of her former grad student colleagues is on her interview commitee. And she is more expert than he is on this topic. Good chance, therefore, that we might move to Europe where I am also a citizen. Give up on the green card quest here. And give up my tenured position here...

Been juggling all kinds of grad school applicants, from China, US, Chile, Pakistan, and Ghana today - getting close to the crunch date when people have to give decisions to universities about whether they are accepting offers. And I'm a graduate studies director. Trying to do deals on scholarships and tuition waivers etc. Some US students will be visiting campus tomorrow. And tomorrow evening I'm going to go visit Snork Maiden in the next state over.

OK, so here is the stuff that is in the title of this post:



All my losses and wins in NQ trading. The size of the losses and wins is on a per contract basis - each trade though might have from 1 to 8 contracts. There is a definite upward trend in the losses, which is good news. You can see the "blow-ups" though. I just need to get rid of them and I will be fabulously successful :P The t-statistic for NQ trades is now 1.87. I figure I've traded about $13 million of underlying value for $6400 in profit at an average gain of $17 per contract... Starting value of the account was $10,000. I also lost $2000 in Australian Dollar and SPX (ES) trades. Wins are at $23.2k for $95 per winning contract on average and losses at $16.8k for $130 in loss per contract. I'm making money because 2/3 of the contracts I trade are profitable.

It's been a tough month for trading but at this point at least I am up on the month and only have 3 QQQQ put contracts outstanding.

Wednesday, March 28, 2007

Emergency Funds

I have a whole list of planned posts on investment performance and asset allocation - but they can wait. This is a great post. Yes, it's all about financial freedom. I think the term "emergency fund" is bad. First it's negative rather than positive. Maybe it works for some people to scare them into saving. But by defining the only purpose for saving outside of retirement accounts and house buying as dealing with emergencies keeps people away from financial freedom and pushes them towards the standard wage/mortgage slave until some far off retirement date. As I wrote in my comment on English Major's post - why do people want to live "paycheck to paycheck"? Of course that's rhetorical... Obviously, some people can't save without the forced mechanisms of retirement schemes and mortgage paydowns. But I think others are brainwashed into thinking this is normal. Not all PF Bloggers fall into this model of behavior but quite a high proportion of them do. This isn't the case among "investment bloggers". They are trying to create financial freedom right now. Of course these aren't mutually exclusive sets - there is a lot of overlap.

Freedom fund is a better term than emergency fund, but once you have a decent amount of money you don't want to be holding it all in cash...

P.S.

I'm now back to being ahead for the month in trading after a couple of good days. When I am in the groove it seems really easy to make money at this. But then there are times when I completely screw up... I just closed out the short I had set up to hold overnight. The oil price suddenly spiked at 4:53pm and stock futures fell. But now the oil price is going back down. Not clear what rumors exactly triggered the huge spike - from $63 to $68 in 8 minutes! I didn't notice that that was the cause of the fall in stock futures until someone on Silicon Investor pointed that out to me. Following on from Sunday's post - this sounds somewhat similar but even less coherent :)

Sunday, March 25, 2007

Asset Allocation: The Big Picture

In response to Finance Girl's request I'm going to do a series of posts on my asset allocation. Today's post will look at how I think about the overall big picture and then the next few posts will drill down to the level of individual investments. Most bloggers out there are either stock pickers, traders, or index investors though there are more complex allocations too. My approach is one of those more complex ones. This is how I think about asset allocation at the highest level:



You're seeing my position when I have almost no active trading positions. Let me explain each category in turn:

Passive Alpha This is a bit of a misnomer - as these funds are invested with active managers - these positions are passive from my point of view because it is largely buy and hold. Included in this category are closed end funds, financial firms (including Berkshire Hathaway and Hudson City Bank Corp for example), REITs etc. The point of these investments is to try to generate returns that are relatively uncorrelated with the stock market and uncorrelated with my own trading. That's why I call these investments "alpha". But as you can see I estimate that these investments actually contribute 0.138 points towards my portfolio's total beta. They aren't pure alpha by any stretch of the imagination. I include the financial firms in this category because really a bank, for example, is involved in investing funds in loans to a diversified portfolio of customers. So in some ways it's similar to a closed end fund. There are also fund management companies in here. In the long-run I'd like to increase the proportion of net worth invested in this category from 22% to 40-50%.

Beta Another misnomer - as all the mutual funds in this category are actually actively managed. So they aren't pure beta. The reason they are all actively managed is something of a legacy of the past - and there still aren't many indexed mutual funds available in Australia. In the future I would probably add some ETFs in this category, but as I want to reduce the allocation to 40-50% of net worth, this won't happen any time soon. You'll notice that this "beta" actually contributes little beta to my portfolio. I adjust my holdings in this category over the course of the four year stock cycle. At the moment I am as conservative as I will get and most of what is in here is in fact bonds rather than stocks. I try to hold these mutual funds for at least a year to get the long-term CGT rate in my taxable accounts. The largest part of these funds though are in my retirement accounts.

Trading This includes cash in my trading accounts and open trading positions. This is where I actively try to generate alpha in my portfolio. My policy is not to add any more money to my trading accounts until I can show that I have recaptured the losses I generated earlier in my trading career. At the beginning of the last week there were about $10,000 of losses to recapture. Now we are back to about $13k. The beta of this category can be either positive or negative. My most extreme positions add or subtract about 0.7 beta to my portfolio - taking me from fully long to net short. Trades are generally directional bets on the market and most are held from 1 day to 1 week, though there are also occasionally longer-term trades.

Industrial Stocks This is the extent of my "stock-picking". Any non-financial individual stock is in here. I don't think it is easy to pick stocks that will outperform (at least not for me) and, therefore, I don't do much of it. If I see something that does look good I'll do it. In the past I had much more allocated to this. In the future, I'll probably have even less.

The final two categories are simpler - Liquidity is cash in non-trading accounts - 3% is around $12,000 - this isn't an emergency fund in any sense though. I use credit cards for emergencies. It's just money I haven't decided yet to allocate to investment or spending. Borrowing - the basic idea is that I am 100% invested and borrowing funds to trade with. You'll see the two categories have about the same allocation. Where possible, though, I try to use products with embedded leverage (futures, options, leveraged mutual funds) rather than borrowing explicitly as the interest rates are generally much lower than I can get myself.

My portfolio is also invested in assets associated with different currencies. In the first few years after I returned to the US I sent a lot of savings back to Australia as the Australian Dollar was cheap. Now that the Aussie Dollar has risen a lot I am accumulating all savings in the US and will begin transferring dividends and distributions I receive in Australia back to the US. In the long-term I would like to bring down my AUD allocation to around 40-50% of my net worth. The other category includes some global closed and mutual funds that are not hedged back into a single currency such as AUD or USD.

For another view of my asset allocation you can check out my NetWorthIQ Profile.

Benchmarking Your Performance

I've been teaching Yoyo and you, the reader, how to compute your investment performance. In today's class we'll compare our results to some broad market indices and get to make a colorful chart. Yoyo definitely prefers pictures to numbers :)

There are two indices I compare my performance to - the MSCI World Index and the S&P500 Index. I use these because one is a broad measure of global stock market performance and the other has traditionally been used as a US market benchmark. But very importantly I have free access to total returns for both indices. This last point is very important. A lot of people out there and especially those in financial publications (like Barrons) compare total returns of a mutual or hedge fund to just the price performance of a stock index. This totally ignores the dividends received by the stocks in the index and dramatically exaggerates the relative performance of the fund in question!

You can download total returns for the S&P 500 here. The monthly total return is the second to last column. You are going to need to construct the total return index yourself as I showed you in the day before yesterday's class. For the MSCI I go to this page to get the latest number for the "All Countries", "Standard", "Gross" World Index. Gross indicates that all pre-tax dividends are reinvested. The option "Net", takes out local taxes from dividends, and "Price" is, well, just the price without dividends. Make sure you also use the number in US Dollars. If you click on the name of the index, you'll be asked to register for free to get to download the historical data.

OK, so I've downloaded the data, computed a total return index for the S&P 500 and come up with the following table:



At the bottom of the table I've given the total return since the end of 2004 and the annualized rate of return, which we learned to compute yesterday. Clearly, being invested in the S&P 500 instead of global stocks resulted in a much lower rate of return over this period (9.8% vs. 16.02% per annum). My returns are somewhere between the two (12.88%).

We still need to do one more thing before graphing these indices - set the values for December 2004 equal to 1000 for all indices. Using Excel we can paste the following formula into a new column:

=1000*b2/b$2

where b2 is the value of Moom's index in Dec 2004. And then paste =1000*b3/b$2 etc. in the cells beneath it to get a new column with Moom's total returns index standardized to start at 1000 in Dec 2004. Does that make sense? We do the same for the MSCI and then we can finally get the pretty picture that Yoyo has been waiting for:



I chose the colors to reflect how hot the performance has been :)

Saturday, March 24, 2007

Computing Your Annual Rate of Return

Once you have your total return index computed it's real easy to compute your annual rate of return for any period you choose. Here is my total return index for the last couple of years:



To compute the annual return for 2006 we just divide the index for December 2006 by the index for December 2005:

1951.8/1642.52 = 1.1883

and subtract one:

1.1883-1 = 0.1883

In percentage notation that's 18.83%.

Of course we can work out the rate of return for any period of more or less than 12 months. From the end of January 2005 to the current date in March 2007 we get: 2088.22/1652.11-1 = 26.40%. We can also work out what that is at annual rate. Here we'll need a program like Excel:

AnnualizedROR = (2088.22/1652.11)^(12/26)-1 = 11.42%

The hat sign is how Excel represents "to the power of". 26 is the number of months from the end of January 2005 till now and 12 is the number of months in a year. We can do the same thing for periods of less than a year... for example my annualized rate of return so far this year is:

AnnualizedROR = (2088.22/1951.8)^(12/3)-1 = 31.03%

If you wanted to compute the average return per month for a given period, you'd just replace 12 with 1 in the formula. Next, I think I'll cover comparing your performance to benchmarks. I'm happy to take suggestions for topics on this theme to cover.

Friday, March 23, 2007

Computing the Total Return Index



Yoyo the green rabbit can help cheer me up this morning :) I get upset when I make stupid trading mistakes, primarily because I see my goal of making a living from trading recede a little bit. After a successful period I'm thinking: "I can do this". And then after a big stupid loss (caused by trading against the model rules) I think: "Maybe I'm not smart enough to do this".... oh well, onto the next step in measuring investment performance.

This step is computing a total return index, or in Australian, an "accumulation index".

Yoyo has computed her monthly returns for three months now:

December 2006: 2.71%
January 2007: 2.28%
February 2007: 2.38%

(Actually those are my returns). The first step in computing the total return index is to choose a base period. Here it is the end of November 2006. We'll give that point in time an arbitrary value, say 1000, though we could choose 100 or any other number, it doesn't make a difference.

Now to compute the value of the index at the end of December we do this calculation:

Index(Dec06) = 1000*(1+0.0271) = 1027.1

And in a general formula:

Index(t) = Index(t-1)*(1+ROR(t))

Make sure that you use 0.0271 as the rate of return and not 2.71! Just divide any percent rate of return by 100 to get the rate of return not in percent. Then:

Index(Jan07) = 1027.1*(1+0.0228) = 1050.52
Index(Feb07) = 1050.52*(1+0.0238) = 1075.52

It's easy to set this up in an Excel spreadsheet. Next up, computing annual and annualized performance.

Thursday, March 22, 2007

Stupidest Trades Ever

I did what must be some of my stupidest trades ever today. I did nothing until the FOMC announcement. Then when the market began to rise when they announced that they were softening their bias towards tightening I went long. Very good up to this point. Then I went short and initially the market fell, but then it soared and I remained short. Despite the model being long! Eventually, I gave up and then started doing some more trades 8 NQ contracts at a time. Initially, I got back part of my loss, but then I lost more.... I blew up about half the profits I have made so far on my IB account - about $2500. The stupidest bit is I know these are bad trades as I am doing them almost... Maybe I just should ignore the market on Fed days like Trader Mike does. Some of my other worst disasters were on Fed days. I lost more than $4000 on that day in June 2006. So maybe I am getting better!

As the Australian Dollar soared after the FOMC announcement I am still above $US400k in net worth. But back below $A500k. My t-statistic for NQ trading (z-score) fell to 1.86. Well, at least I am above the low after my previous blow-up. And the t-statistic is way above the level it was when I first computed it. So I am making progress of sorts.

Here is a chart showing today's trades:

Contributions of Investment Returns and Savings



Following up from trainee investor's question on my previous post I made a jpeg of the Excel Chart that I maintain on the relative contributions of savings and investment returns to my net worth growth. Yoyo is patiently waiting for her next class on measuring investment performance :) Clearly, most of the current growth is coming from investment earnings, while in 2005 most growth came from savings. Back in January 2003 my accumulated investment earnings were negative to the tune of -$23,944. Now we are at a positive $173,768. So I have gained almost $198k. In the same period I've saved $142k. So there you have it. By the way the investment returns are primarily pre-tax (the exception is my Australian Superannuation Fund where the 15% tax on earnings is deducted at source). By the way any tax payments would be recorded as negative saving. But so far I have only had tax refunds as my withholding from salary has been sufficient to cover everything.

Wednesday, March 21, 2007

$400,000

Yes my net worth is now above $400,000. But there are no guarantees that I'll still be here at the end of the month for the reasons I discussed in my post about exceeding half a million Australian Dollars. Nice to see that 4 at the beginning of the number :)

Measuring Leveraged Investment Performance

Following up on Enough Wealth's comments on my previous post, I'm going to do an example here where Yoyo has borrowed money to invest. This could apply for either stocks bought on margin or an investment property bought with a mortgage.

31 January 2007:
Brokerage - stocks and cash: $45,500
Brokerage margin loan: -$10,000
Retirement: $93,000
Total: $128,500

28 February 2007:
Brokerage - stocks and cash: $45,500
Brokerage - margin loan: -$12,100
Retirement: $94,200
Total: $127,600

Yoyo computes the value of her accounts by adding up all the assets and deducting the value of the margin loan. If she invested with Ameritrade for example, she wouldn't even need to do this as they give you the net value of your account already computed. I am assuming that she pays $100 in interest and borrows an additional $2000. It's usual to capitalize all these costs onto a margin loan, so she doesn't even need to do the calculations.

She also collects the following information as before:

February retirement contributions: $550
15 February put $500 into brokerage account
28 February withdrew $600

The calculation proceeds as before:

Investment gain in brokerage account = ($45,500-$12,100)-($45,500-$10,000)-$500+$600 = -$2000
or: monthly change in account value - net contribution to account.

The calculation for the retirement account is unchanged - a gain of $700. Yoyo's net loss is now -$1300.

Her rate of return is now:

ROR = -1300/(45500-10000+93000) = -1.01%

If your broker reports the net account value like Ameritrade does then there is no additional calculation at all compared to the non-margin case.

For an investment property bought on a mortgage, all payments towards paying the mortgage count as contributions. Payments for maintenance, insurance etc. also count as contributions to the account. All my financial investment examples compute pre-tax return so I would not include payments for taxes in the real estate calculation either in order to derive the pre-tax return.

The main point here is that borrowing money or paying off a margin loan by selling stock can be completely ignored when you are calculating return on investment.

One final point - the calculations EnoughWealth gives are annual rates of return. I recommend to first compute monthly rates of return. My examples so far are all for monthly rates of return and are not annualized.

Tuesday, March 20, 2007

Getting Started in Measuring Investment Performance

I've been posting a lot of different analyses of investment performance recently. They all rely on the same underlying data which are in fact fairly easy to compute. I have the impression that people think it is harder to compute that data than it is. Several trading blogs I read report daily trading results and several PF bloggers report some annual investment performance figures, but I haven't seen anyone else compute the kind of statistics that I've been discussing here. Let me know if you know other blogs that are computing longer term investment performance statistics.

What you need to know is the value of the accounts you want to include in your computation at the end of each month and how much money you put into and out of the accounts. I use my total net worth and calculate how much I earn from salary etc. and how much I spend each month. But you can look at just investing and trading accounts and ignore checking accounts, credit cards etc.

OK, so here is a simple example. Let's assume that Yoyo has a brokerage account and a retirement account. Yoyo checks up her statements for the two accounts for the last couple of months:

31 January 2007:
Brokerage: $35,500
Retirement: $93,000
Total: $128,500

28 February 2007:
Brokerage: $34,400
Retirement: $94,200
Total: $128,600

She also collects the following information:

February retirement contributions: $550
15 February put $500 into brokerage account
28 February withdrew $600

That's all the data needed. Now let's help Yoyo do the calculations for February's rate of return.

Investment gain in brokerage account = $34,400-$35,500-$500+$600 = -$1000
or: monthly change in account value - net contribution to account.

And for the retirement account:
$94,200-$93,000-$500 = $700

So Yoyo's loss for the month is the sum of these two investment returns = $700-$1000 = -$300.

To get the rate of return we divide this number by the value of her accounts at the beginning of the month:
ROR = -300/128,500 = -0.23%

The hard bit is doing this for every month you want to include in your statistics. I'll explain some basic statistics in an upcoming post. Please let me know if anything is unclear here!

Aussie Dollar Trading Above 80 US Cents

It hasn't been above 80 cents in a long while. My Aussie Dollar model gave a sell signal yesterday. It's good I'm not trading forex anymore as I would have shorted the AUD and probably gone long the Yen, which would have made things twice as bad! This model is a previous generation model that is not as precise at all as my NDX trading model. One day maybe I will apply this new model to forex. But in the meantime I've decided to take a slow and steady approach to foreign currency management. Over time I plan to move my portfolio towards 50:50 Australian and US Dollars by accumulating savings in the US and transferring dividends and distributions from Aus to the US. After reaching the 50:50 level I would do regular quarterly, or whatever, rebalancing. Currently I am approximately 67% in AUD, 26% USD, and 7% other (global mutual funds that are unhedged or hedged into other currencies).

BTW I just shorted 2 NQ contracts @ 1781.75. The model is still not giving a very clearcut signal, so this is a hunch based on a bunch of stuff. I'll see in the morning whether I should add to it. Maybe this evening already I'll decide to get out of it if there is a very strong rally in Asia.

Closed Positions Too Soon

Looked this morning to me that the major indices were going to try to close the opening gap. So I closed all my long trading positions. I got to campus now and the indices are up nicely. As I read recently: "It's better to be out of the market wishing you were in than in the market wishing you were out" :) This is especially true if your method generates a high percentage of winning trades. Then it is less important to let the winners run. The model signals are very ambiguous here, so I'll wait for a clearer signal to do that. I'll do a post or two soon on the basics of computing your investment and trading performance. I think it is much easier than most people think it is.

Monday, March 19, 2007

Time-Varying Alpha and Beta



When estimating alpha and beta, a big question is how much data to use? 36 months? 60 months? Different periods will result in different estimates and those estimates will change as new data comes in. Advanced time series econometrics methods can use all the data available to estimate alpha and beta coefficients that vary over time. To produce the chart above I used the Diffuse Kalman Filter to estimate alpha and beta. I assumed that alpha and beta could each be modeled by what is called a local linear trend model.

Alpha increases smoothly over time. The model couldn't distinguish any changes in the slope of this learning curve. With more data over time it might be able to. It could distinguish a lot of variation in beta. I increased my beta from when I started investing from around 0.35 to a maximum of 1.35 in September 2001. Increasing my beta in this period after March 2000 was not a good idea! In recent years beta has fallen to its current level of 0.42. Alpha is currently 10.3%. The expected rate of return is, therefore, 17.6%.

We can also use this model and the actual monthly returns to the MSCI Index and the risk free rate to compare actual returns to those that are predicted by the model and the MSCI index:



Deviations from the predictions are the "noise" that is not explained by the MSCI and the risk-free rate or the smoothly trending estimate of alpha.

Sunday, March 18, 2007

Expected Returns

Following on from my previous post, one of the uses that this alpha/beta analysis can be put to is estimating your future expected rate of return. You just need to plug in the expected risk free rate and benchmark rate of return into the following formula:



The "hats" indicated "estimated value". This formula ignores the uncertainty in the estimates of the coefficients. You could do a Monte Carlo simulation to find the expected distribution of future returns. But for now I will assume that the coefficients are known with certainty. Based on my MSCI analysis I come up with:

alpha = 11.26%
beta = 0.72

For the S&P 500 I get:

alpha = 16.07%
beta = 0.58

(see what I mean by the MSCI being a steeper hurdle?). Now if I assume that F = 4.98% (its current value) and M = 10.5% I come up with expected rates of return of 20.2% and 20.1% based on the two benchmarks. Of course, you can assume a lower rate for M if you want. How do these compare with my actual returns?



All these rates are annualized. For the last 12 months and the last 3 years my annualized rates of return are 23%. However, 2005 was a bad year and so for the last two years the mean is just 14%. The MSCI achieved 11.6%, 14.4%, and 15.9% for the same periods. So the index was a bit above trend in those years and my own results for 1 or 3 years are also above trend. The S&P 500 has been at or below trend, however. Over the last 5 years I averaged 13.6% compared to 10.7% and 5.8%. My alpha has been increasing over time so this makes sense. Over a ten year horizon I underperformed the indices.

Let's assume that the 20% rate makes sense, then what does it imply? 20% of my $215k in non-retirement accounts is $43,000 roughly the US average salary. It is enough pre-tax income to cover my current expenses. I could "retire" now. The quotation marks mean that I would become a full-time active investor and trader. My retirement accounts would continue to grow at a rapid rate. Real retirement could happen at a later date if I wanted it. Of course, I would like to have a huge margin of safety so that if the rate of return is lower or my expenses rise rapidly I wouldn't run out of non-retirement money. The bottom line is though that all non-retirement saving I do from now on is increasing that margin of safety.

Saturday, March 17, 2007

What Does Alpha Look Like?

We hear a lot about "alpha" - the risk-adjusted average rate of return for an investment portfolio such as a mutual or hedge fund. Technically, alpha is the coefficient in the following regression equation:



R is the rate of return on the portfolio measured in each period t, F is the risk free rate of return (return on a 90 day U.S. Treasury Bill), and M is the rate of return on the market portfolio, which is often approximated by the S&P 500 total return index or the MSCI World Index etc. Usually monthly data is used to compute the series of returns - typically 36 to 60 months of data. Alpha and beta are then the coefficients of the ordinary least squares regression of R-F on M-F. You can do this in Excel using the function LINEST.

Beta shows the sensitivity of the portfolio to the market rate of the return. Alpha shows the average of the returns that are not explained by the market rate of return. The last term epsilon is a series of error terms that represents the variations in return not explained by the market rate of return - the noise in the regression.

But what does alpha actually look like? The following chart shows a scatter of my rates of return for the last 36 months against the returns of the MSCI All Country Gross World Index:



The trendline computed by Excel shows the part of returns explained by the MSCI index - its slope is beta. Alpha is the height of the trendline above the zero, zero point on the chart (I haven't deducted the risk free return here by the way). This is a monthly alpha - the annual alpha here is something like 12%, which compared to most mutual funds is rather high. But I am mostly struck by how fragile alpha looks when surrounded by the wild monthly fluctuations in the portfolio returns.

Thursday, March 15, 2007

Don't Worry About Missing Out on Big Gains

Worrying about missing out on big gains often leads to foolish actions. Whether it is first time home buyers worrying about being "priced out of the market forever" or me a couple of days ago thinking I wasn't taking big enough positions. This led to me breaking my trading rules yesterday - taking on positions in the middle of the day that I wasn't intending to just turn around for a quick scalp but instead hold overnight. My rules say to close trades in the morning and start opening new ones towards the end of the trading day. I kept trying to go short and the market kept rising against me. And I missed the best opportunity to cover my short positions at a profit earlier in the day. I'll close out at least half my position later this morning whether it is profitable or not. The economic releases at 8:30am this morning have pushed the futures into the red from the green so the damage is at least not going to be as bad as it seemed it would at one point I think.

My current goal for trading is to demonstrate that I can consistently generate some level of profits. I don't need to capture all the market movement. But I keep wanting to hedge my investment portfolio against adverse market movements, though that is not my goal for trading - the long-term goal is to generate returns that are uncorrelated with the market (over periods of a month or more - obviously they are highly correlated in the short run as I am taken directional bets on the market). I'm almost half way to my annual goal of $19k in trading profits. I don't need to take big risks as I am earning a salary, I don't depend on trading for a living.

P.S. Great article on the economics of homeownership

Increasing My Position

I shorted 800 QQQQ @ 42.62 and 2 NQ contracts @ 1752.75. Now I am short the equivalent of 3200 QQQQ shares and have a negative portfolio beta overall (similar to being net short overall but I take into account the sensitivity of each financial instrument to movements in the S&P500) of -0.18. Let's see if this works.

Wednesday, March 14, 2007

Right Idea, Insufficient Followthrough

I closed out the short NQ futures contract just before 12 today as I was going to go teach my class and the market didn't seem to be going anywhere. Then the market collapsed of course while I was teaching. It looks to me like this is the beginning of the next wave down to match the initial one that peaked on February 27. At this point though I only have 3 QQQQ put options and a short position of 500 QQQQ shares. Between them they contribute beta of -0.15 to my portfolio. My overall portfolio beta (to the S&P 500) is 0.25. I should have more confidence but after the trading break I was stepping into the waters again nice and slowly. I'm up $1580 in short-term trading this month at this point, but could be doing a lot better.

Tuesday, March 13, 2007

Getting Back into Trading

Last week I took a complete break from trading while we were in Florida and after we returned. Today I initiated a small position. This morning I used half the cash in my Roth IRA to buy 3 April $46 QQQQ puts and I also shorted 500 QQQQ shares in my regular trading account. This is a minimal position equivalent to 1 NQ contract. Only one of my indicators was on short today. This afternoon the market rallied and I shorted 1 NQ contract in addition. So now I am at 1/3 of my maximum position. Two more of my indicators are close to a short tomorrow. There are retail numbers out in the morning at 8:30am. If the market does turn to the downside this could be the second downwave of the correction. Or it could just be a wiggle inside a sideways correction of the first downwave. I don't really need to form an opinion about that. I will just keep on building up my position as things get more bearish or vice versa. But the downmove should last way beyond the end of this week based on the forward forecast. Of course if the stochastic some how manages to get above 80 the short signal will be negated.

Monday, March 12, 2007

Has a Bear Market Started?



The chart maps out a possible Elliott Wave interpretation of the move up since last summer. It's pretty clear if you look at the waves and technical indicators that the current correction is unlikely to be complete. We should expect at least another downwave comparable to the one that started in late February. The real question is: Would that be the end of the correction or has a more extensive bear market started? I've been looking for a 20% correction in the stock market to end the rally that started in late 2002, which has turned out to be one of the longest periods in stock market history without a 20% correction.

From a fundamental perspective there are plenty of signs of weakness in the US and global economy but nothing yet is pointing to a severe recession rather than a mid-cycle slowdown of the type we saw in the mid-80s and mid-90s. Even if there is a recession, recent US recessions have been very mild by historical standards. There is a trend to less and less volatility in the business cycle over time - with longer and longer expansions and milder and milder booms and busts. Looking at stock valuations, they seem fairly reasonable (though not cheap) given the low interest rate environment. Profits are at an all time high as a share of GDP in the US and have increased by more than ever in percentage terms since the last US recession. If a recession is coming profits will fall and so should stock prices but the Fed is likely to cut interest rates and inflation is contained. I just can't see an argument for a very steep fall in stock prices. I tend to be bearish but I am not a gloom and doom gold-bug type permabear. The economy does change over time and the past doesnt repeat itself in exactly the same way. We are in the middle of a massive global economic integration on a scale not seen since the late 19th century. Manufacturing is a smaller share of employment and GDP (but not of material production neccessarily - huge technological change has resulted in the price of manufactures falling relative to everything else) and governments and central banks know better how to control the economy than in the past.

From a technical perspective it depends where the wave in the chart above fits in the bigger cycle of things. It is rather short in time and amplitude compared to the rally wave that started in 2002 and maybe ended in late 2004. On the NASDAQ indices there was then clearly a correction lasting till July 2006 and then this current rally that has just ended. The bullish interpretation is that from 2002 to 2004 was wave 1 of the post dot.com bust bull market. Wave 2 played out in 2005-6 and the recent rally is the first wave of wave 3. Our current correction is then the second wave of wave 3 and could complete in a matter of weeks. Bearish e-wave scenarios look a lot less convincing in the US markets. The most obvious is that the recent rally was wave C of a corrective wave starting in 2002, which itself is part of a decade long triangle correction that started in 2000. But as I pointed out above this C wave would be rather small. Most non-US markets have clearly been in a bull market over these years. The weakness of the bull-case is that if the current correction does not take out last summer's low in the S&P 500 we will not see a 20% correction. Then the bull market that started in 2002 will turn out to be the longest in history. Are things really different this time?

Sunday, March 11, 2007

Investment Decisions for my Mom and More Performance Statistics

I've been reviewing the information a new manager we hired for my Mom sent me about potential investments. He has a small firm that invests in the US based in the country where she lives. We have so far given his firm about 9% of her assets. At the moment the money is all in a money market account. But now he wants to know whether after the recent small decline in the markets we want to invest in stocks. I'm going to recommend to do so but to start with $100k and then invest $8-$10k each month over the next year in the same investment. I think it is likely that the stock market will decline further this year, but I can't be sure. As I don't think stock valuations are excessive, the Fed is likely to cut interest rates, and inflation is contained, I don't think any stock market decline will be pronounced. Therefore, I'm recommending this dollar cost averaging (DCA) approach. She has plenty of cash and bonds in other accounts and only has about a 15% allocation to long-only stocks at this point. When this investment is complete she'll have about 25%. I am also recommending to invest with Aletheia Research and Management. This is the kind of manager I love - their alpha is 15% or so though the growth account also has a high beta. It isn't the kind of investment your typical pf-blogger likes as the expense ratio (paid as a "wrap fee") is rather outrageous (3% I think). The minimum investment is $100k which is why we can't just DCA all the way.

While I was reading the information on these separately managed accounts I came across an interesting performance statistic - down capture ratio and up capture ratio. These statistics are the fraction of the market gain a manager captures in months when the market rises and vice versa. It is easy to compute in an Excel spreadsheet. The Aletheia Large Cap Growth Managed Account reports an up ratio of 1.65 and a down ratio of 0.47. This is a very impressive asymmetry. According to Alexander Ineichen, asymmetric returns are the hallmark of hedge-fund like performance. So of course I computed my ratios for the last 36 months: 1.22 up and 0.71 down. Not bad. It wasn't always that way :) The asymmetry is increasing over time. In my early days of investing the asymmetry went in the opposite direction!



Following my discussion yesterday with Rich Gates and BackOfficeMonkey (love the handle!) I took another look at the correlations of my returns (since October 2002) with several other assets and managers. As I knew already, my largest correlation is with the MSCI World Index. I have a small negative correlation with the TIAA Bond Market Fund and a small positive correlation with the Australian Dollar. I estimated a regression against all these factors and only the beta on the MSCI World Index was statistically significant. If I had the data to hand I'd do a more sophisticated analysis, but the point is that the MSCI is not a bad benchmark.

Saturday, March 10, 2007

Timing and Measuring Investment Performance



The chart shows how many percent you would be ahead of the MSCI World Index (Gross = total return index) today by investing with me in each of the months on the X axis. This shows the influence of timing of investments on performance. Measured from inception in October 1996 when I first started measuring investment performance you would be 10% behind the index. Investments made in late 2004, through 2005 would also be lagging in performance. But if you invested with me in any of the other periods you would be ahead by up to 47%. If, instead, one dollar cost averaged into this investment program - as I have done in practice - you would be beating the market, as investing in most months results in index-beating performance.

I think a chart of this sort would be very useful as part of every mutual and hedge fund prospectus or as a feature of websites such as Yahoo Finance.

Friday, March 09, 2007

February Report

All figures are in US Dollars unless otherwise stated. Income and Expenditure

Expenditure was $1771 - 54% of take home pay ($3,299). This is the normal background level of expenditure. 403b contributions again totaled $1,795 and Roth contributions $333.33. Non-retirement investment returns were quite strong again this month ($5,777). Retirement investment returns were also nicely positive ($3,202). The rise in the Australian Dollar contributed significantly to returns.



Net Worth Performance

Net worth rose by $US12,301 to $US389,704 and in Australian Dollars gained $A8,017 to $A493,859. This is $US7,000 more than needed to be on track for my 2007 goal of a net worth of $US470k. The Australian Dollar rose this month resulting in a relatively large gap between performance in the two currencies. Non-retirement accounts reached $US212,618 or $A269,444. Retirement accounts also saw nice gains to $US177,086.

Investment Performance
Investment return in US Dollars was 2.38% vs. a 0.49% fall in the MSCI (Gross) World Index, which I use as my overall benchmark and a 1.96% fall in the S&P 500 total return index. Non-retirement accounts gained 2.81%. Returns in Australian Dollars terms were 0.78% and 1.21%. U.S. Dollar returns also beat the indices over the last 12 months:



The contributions of the different investments and trades is as follows:



The returns on all the individual investments are net of foreign exchange movements. Foreign currency gains appear at the bottom of the table together with the sum of all other investment income and expenses - mainly margin interest. QQQQ/NQ trading again yielded very strong returns ($4,026). Apart from trading and foreign currency diversification the monthly return would have been flat. The most positive contributions from investments were from Everest Brown and Babcock - an Australian listed fund of hedge funds and hedge fund management company - and from two bond or balanced mutual funds. If all my investments were in US stocks I would have had a negative month. Diversification across currencies, asset classes, and trading vs. investing generated the positive return.

Progress on Trading Goal
Trading in my US accounts netted $3,826 a 13.9% return on trading capital. The model gained 9.4% while the NDX fell 1.7%. My goal for the year is to end up with at least as much in my three accounts - regular trading, Roth IRA, and IB - as I've put into them. The accounts in total gained a net $3,374 and I have now achieved $7,702 of the annual goal of about $19,000. Since the beginning of the year the trading capital gained 36% while the NDX was essentially flat.

Asset Allocation
At the end of the month the portfolio had a beta of 0.40. 40% of the portfolio was in stocks, 43% in bonds, 13% in cash, and loans totalled -7%. The remainder was in hedge fund type and real estate investments, futures value etc.

Half a Million

Australian Dollars. I just updated my net worth for the current date in March and I am above half a million Aussie Dollars. I might not be able to hold on to that level as some day soon I am going to reduce the carrying value for Croesus Mining. But only when it either delists or trading recommences on the Australian Stock Exchange. I expect that will be a $A10k-11k hit to net worth. US Dollar net worth is currently down on last month's strong close due to the fall in the Aussie in recent days.

It is just under a year since I passed the last Aussie Dollar milestone.. $A300k was first passed in January 2005 and $A200k in June 2004. But I first exceeded $A100k in December 1998:



Superannuation refers to all retirement accounts and medium term balance to non-retirement accounts.

Got back last night rather delayed from Florida to the snow and ice of the northeast. Today has just been catching up on everything and otherwise being lazy.

Friday, March 02, 2007

Preliminary Report

Initial figures show an almost $11k gain in net worth to $388k and well ahead of the target of $382k for this month and an investment performance of 2.05% mostly due to the gain in the Aussie Dollar as well as trading gains, my mutual funds were up just a little. There will be a full and detailed report with the final figures after we get back from Florida. Today a hedge trade went wrong when the market swooned this morning but I managed to come out with a profit on both sides of the trade.... eventually.

The main driver of the fall in the stock market seems to be the rise in the Japanese Yen. People who have borrowed or shorted low interest Yen to use the proceeds to invest elsewhere are covering their positions as the Yen rises and liquidating other investments elsewhere. The media calls this "unwinding carry trades". The fall in the Chinese market and the softening signs in the US economy also are playing a role of course.

Also, check out the latest 30s and 40s carnival.

Thursday, March 01, 2007

The Day After

At this point it looks like the market is shaping up for a sideways day after the rebound in Shanghai overnight and smaller losses in European trading than Wall Street experienced yesterday. I put on a hedge trade - long 2 ES and short 2 NQ which is biased to the long side, but exited it at a loss as I had a meeting at 9:30 and it didn't seem to be doing well. It would have been in the money by now. So I guess I am just jittery after yesterday though I made plenty of money in the US market, but got slammed of course in Australia. I don't know the full extent of damage there as mutual fund prices are reported with a delay. Worst are probably the losses in resource stocks affecting my CFS Global Resources Fund which has about 10% in each of RIO, BHP, and RTP (those are the US tickers). The more hedge fund like instruments mostly responded better, though EBI.AX fell 5% and hedge fund manager EBB.AX fell 8%. I reckon I was probably slightly up yesterday overall, though I don't know how the short-term Australian Bonds in the CFS Conservative Fund (my biggest holding) responded. The Aussie Dollar fell and US bonds rose.

But I'm not worried that I don't have a trade on. I don't have to have a trading position at all times. I used to think of my short-trades as hedging my long positions and felt nervous if I didn't have a short when I thought the stock market might fall. But now I think in a much more "alpha-centric" way. Alpha are returns that are not correlated with stock market returns while beta reflects returns that are correlated with the market. I divide my portfolio into three sections:

1. Hedge fund type instruments that hopefully generate alpha. Sometimes some of these seem to have a bit more beta than I was reckoning on.

2. Beta - mainly long-only mutual funds whose return is mostly correlated with market returns and may have a positive alpha. It's no big deal if their alpha is negative to a small degree (one reason I am not worried about expense ratios which tend to reduce alpha) because I can generate alpha elsewhere in my portfolio. I change my exposure to these funds over the 4 year stock cycle. At the beginning of the cycle my exposure to stocks would be much bigger. I don't have to get my market timing perfectly right. At the moment I am 50-50 in bonds and stocks reflecting the late stage of the cycle which maybe now is heading towards the bottom - assuming that we need to see a 20% correction before the cycle is over. At the beginning of the cycle I will use leveraged stock funds and margin.

3. Short-term trading - I regard this now also as a generator of alpha. The trades are in ETFs or futures and so have a +1 or -1 correlation to the market while the trades are on. But the stochastic model has a zero beta coefficient to the NASDAQ 100 index. So in the long-run the returns are all pure alpha.

And of course I am also diversifed across US and Australian Dollars.

I've arrived at this strategy after a lot of experience and seeing what works and what does not and what I can tolerate emotionally. It's much more sophisticated than the standard "buy and hold" long-only models. I know I can't tolerate the fluctuations that that leads to. It was interesting seeing the responses of some newbie investors yesterday to the drop in the markets - which in the Dow was significant for a one day drop but was really not that much of a decline off the highs yet. I wonder how many will throw the towel in when we are down 20%? I started investing and trading in 1997 and have been through the high volatility of the 1998-2002 period. I also remember very well the crash in 1987 though I wasn't invested (I did buy a little in some Israeli mutual funds before then and was an undergrad economics and geography student) and even dim memories of the 1970s. Even then I was interested in investing and would discuss things occasionally with my father who was a long-time stock and mutual fund investor though he certainly wasn't wealthy then (we were definitely lower middle class in Britain) and read the financial news.

Anyway, here is what happened in one account, my account with Interactive Brokers:



You can see the big dip a couple of weeks back and then yesterday's recovery, followed by more erratic trading. My Ameritrade account would have a similar pattern. My z-score in NQ trading (total of 209 contracts traded) is now 2.03, which means that the probability that my performance is random or actually negative is something around 2%. The Kelly ratios though for both this and my weekly results in my IB account say that I should be taking on huge amounts of leverage. The Sharpe ratio for the weekly returns on my IB account is 1.68, which is a respectable number for a hedge fund.