Optimal Portfolios

I have been doing some experimentation with designing optimal portfolios, something which I last looked at in 2011. I have the monthy rates of return on various asset classes going back to 1996. These include international shares (MSCI World Index, gross) both hedged into Australian Dollars and not. Australian shares (ASX 200 accumulation), Managed Futures (a mix of Man AHL and Winton), direct real estate (a particular US fund as a proxy), hedge funds (HFRI index), the bond market (again I'm using a fund as a proxy), Australian Dollar cash, and gold in Australian dollars. You can use the solver in Excel to find the allocation that monthly rebalanced gives the highest Sharpe Ratio. This optimal portfolio varies over time but generally it doesn't like hedge funds and allocates about 10-20% to gold, and 20-40% to managed futures. Because future performance won't necessarily be the same as past performance (particularly a worry for managed futures) and because managed futures, in particular, are not tax effective – they pay most income out subject to marginal tax rates – I wouldn't allocate according to a particular optimization. A target portfolio gets near the optimal performance while being more diversified and a bit more tax effective:

This graph shows the performance of various assets and a "target portfolio":

Here the target portfolio is 25% international shares (half hedged into Australian dollar and half not), 25% Australian shares, 25% managed futures, and then 5% in each of real estate, bonds, cash, gold, and hedge funds. Then the whole thing is geared up a bit with borrowing. It performs pretty nicely over various historical periods.

Here we have a close up of performance since the financial crisis:

I've managed to match the performance of the Australian index but have lagged behind the MSCI World Index. It matches the performance of the MSCI but has a smoother path. The next graph shows ten year rolling returns:

Here we see that such a portfolio clearly dominates in the long-run over regular stock indices or my own performance, which has not been good over a ten year period recently. The graph also shows how the performance of the Australian stock market has declined. It had very high ten year  returns prior to the crisis, but now has lower returns than international shares over the last ten years.

I have been moving in the direction of the optimal portfolio by diversifying out of Australian shares and buying managed futures, but it has been too slow so far. In the last few months I have been buying $A10k of managed futures each month. I also allocated more to international investments when I reinvested my CFS superannuation fund in their wholesale funds.

ASX200 Alpha and Beta

Another new chart:

This is based on regressing my returns in excess of the RBA cash rate on the ASX200 returns in excess of the cash rate using 36 months of data. Clearly there is a negative correlation between alpha and beta. In recent years beta is less than one and alpha greater than one. Alpha was very negative during the financial crisis and there are some wild swings before that. The tech crash also had hugely negative alpha. Looks like I outperform in bull markets and underperform in bear markets but that it isn't all just due to too much leverage.

UBS are Recommending 34% Allocation to US Fixed Income - Really?

UBS recommending 34% allocation to US fixed income. I guess this might makes sense if they mean treasury bills (90 day maturity). Don't pay any interest (but not negative like some places in Europe) but US Dollar might still appreciate. Longer term US bonds seem risky if interest rates will eventually go up. I wouldn't rule out though us being in a new long-term zero risk free rate equilibrium. I suppose that this allocation was intended for US clients?

My Mom's money managed now fully by UBS (but she is near their minimum entry level net worth, not what they think of as wealthy) is mostly in fixed income now due to the court order we got. Actually, it looks like that there are no US government bonds or corporate bonds in her account at all, though they are all US Dollar bonds. Things like World Bank, Province of Ontario, EBRD, African Development Bank, Statoil, Shell, Swedish Export Credit Corp etc.

Predicting Stock Market Returns

This correlation is pretty amazing. It is also pretty clear that there is causation here too or both variables are driven by the same other causal variables.

Second Investment in Managed Futures

I have long seen the advantages of managed futures funds. The best of managed futures funds companies seems to be Winton. I previously made an investment with Man-AHL. The fund hasn't made much money for us, but did much better in the financial crisis than most of my other investments. We have 0.80% of net worth invested in the fund. We also have some investment in commodities via GTAA. Another fund that hasn't done much of anything so far. Now I have made an initial investment in a Winton fund offering. The investment is 4.6% of net worth. This takes exposure to commodities out of net worth to 6.0% and out of gross assets 4.5%. The main downside to this fund is that in Australia it doesn't have any tax advantages compared to stocks, which have strong advantages. This means that this will likely remain a small diversifying investment until maybe one day I set up a self-managed super fund, which is a tax advantaged structure itself.

How does this fit into our overall investment strategy? Basically we have a 60/40 portfolio with 60% in stocks and 40% in other investments. Within the stocks 2/3 are planned to be Australian stocks and 1/3 foreign. Within those categories we also allocate to large and small cap Australian and to US and non-US stocks in proportion to their market capitalizations. In the 40% other we have allocations to: bonds, real estate, hedge funds, commodities, private equity, cash, and other. The whole portfolio is then levered to provide about a beta of 1 to the stock market and rebalanced on an ongoing basis. The leverage of a diversified portfolio is an idea from the risk parity approach. 60/40 is simply the traditional stock-bond ratio used for diversified portfolios, and we weight heavily to Australian stocks for tax reasons. Several of the supposedly non-stock investments are in fact Australian listed stocks that are listed investment companies pursuing alternative investment strategies. A lot of the leverage is obtained by investing in leveraged (geared) managed stock funds rather than using margin loans ourselves. We keep the actual margin loan quite small most of the time. This is because the interest rate we can get is much worse than what the funds can get. Interactive Brokers has much better interest rates, but they aren't giving loans to Australian investors at the moment. All this seems to me a reasonable strategy for a non-high net worth investor based in Australia.

Investment Returns Since the Financial Crisis

Following up on yesterday's post comparing global, US, and Australian index rates of return and my own investment performance, today's post looks at the period after the low in the stockmarkets in March 2009. I also compare everything properly in both USD and AUD terms. First, in US Dollar terms:

Initially I tracked the ASX200 very closely coming out of the low but then gradually got pulled down by the lower performance of foreign shares. In the last year the ASX has underperformed foreign shares and that too seems to drag on our performance. Seem to be getting the worst of both worlds rather than the best! In AUD terms things look a bit different (just use the AUD/USD exchange rate to convert everything to AUD values):

Foreign shares performed poorly coming out of the financial crisis. Since mid 2011 they have outperformed Australian shares and as a result I've underperformed the foreign indices in that period.

10 Year Rates of Return: ASX200 vs MSCI, SPX, and Moom

Another way of showing just how extraordinary the performance of the Australian stockmarket has been in the last couple of decades. I've posted this graph before, but now I've added the rate of return of the ASX200 index, which is the 200 largest stocks by capitalization on the Australian share market:

Returns are the average annual rate of return over the 10 years previous to the date marked. The ASX almost hit 15% per annum over ten years at one point! This maybe isn't a fair comparison as Moom, MSCI, and S&P500 are returns in US Dollars and the ASX200 is the return in Australian Dollars. ">Again you can see that I track the MSCI pretty closely.

Investment Performance Against 4 Different Indices

BigChrisB sent me the ASX200 Accumulation Index data he had collected and I have now measured my investment performance since 1996 (monthly data) against it and compared that to the other indices I've been tracking performance against:

The table shows that you get very different performance figures depending on which index you benchmark against. First the MSCI World Index in USD terms and using the US risk free rate to do the standard CAPM regression analysis. In other words, I measure my investment returns in US Dollars too. Estimated beta is 1.23 - a 1% change in the index is associated with a 1.23% change in my portfolio. Alpha is 0.44% which means I am beating the index on a risk adjusted basis. My monthly percentage rate of return is most correlated with the returns of this index. R-Squared is 0.74 which means that 74% of the variation in my rate of return is explained by the changes in the index. Results are quite different when I measure my rate of return and that of the index in Australian Dollars. The R-Squared is only 0.39, beta is much lower, and alpha is a little negative. Switching back to US Dollars my correlation with the S&P 500 is worse than with the MSCI but I underperformed the index by 1.6% per year, risk adjusted. Finally, in comparison to the Australian ASX 200 index and measuring things in Australian Dollars I underperformed by 3.89% a year, beta is 0.89 and R-Squared is 0.51. The ASX has been a fantastic performer over this period of time:

This explains why I benchmark against the MSCI in USD terms even though I live in Australia.

Sources of Data on Stock Market Accumulation Indices

It's easy to get real time information on a wide range of stock market indices and historical data on a daily basis from sites like Yahoo Finance. But getting data on accumulation indices is much harder. Accumulation indices or in US terminology "total return indexes" include dividends in the index return. So they track how much your money would be worth if you could invest in the index and then reinvest all dividends without paying taxes or management fees. I have collected data on the MSCI All Country World Index and S&P500 total return index going back to 1996. My current sources for these indices are here and here.

The MSCI site allows you to select different pages of index returns using menus. I use the ACWI for the Developed and Developing Country, Gross returns, and Standard (midcap and large cap) options. Gross returns also include tax credits. I would include small caps, but that data wasn't available when I started following the series. You can select the data for any date you like.

At the S&P website you need to get the daily fact sheet on each index which includes price only and total return indices. The only way to get exact end of calendar month data is to collect it on the last day of each month. It seems that you now need to subscribe to get historical data. They used to provide a few months for free. You can get Australian indices here as well as the US ones I usually collect from here.

Bigchrisb recently linked to another source for the Australian ASX200 accumulation index - statistics provided by the Reserve Bank of Australia. You can get monthly data for a few recent years there.

Moominvalley August 2013 Report

We have been back home all this month and spending modestly by recent standards. This month's accounts in US Dollars, as usual:

Retirement contributions were higher than normal at $3,630 (AUD 4079) because Snork Maiden got an extra retirement contributions in after last month's three paychecks. Hmmm... I must be missing a contribution somewhere. But there is in fact little line between three paycheck months and three retirement contribution months as retirement contributions don't seem to be sent to the fund on as precise a timing as employees get their cash payment. Snork Maiden's pay was also a bit higher than expected. I haven't checked why.

The Australian Dollar fell a little more this month to exactly 89 US cents. Investment returns in US Dollar terms were 0.68% but in Australian Dollar terms they were 1.34%. For a change, we did much better than global markets due to our focus on Australia. The MSCI lost 2.04% (USD terms) and the S&P 500 2.90%.

Net worth rose in US Dollars by $17k to $873k. In Australian Dollar terms it rose by $A25k to $A981k. Unless there is a sharp fall in the markets we would clear 1 million Australian Dollars this month or next. But this time of year is often volatile :) September has been the second worst month since I started investing (May is worst), while October has been the best. That's surprising given its reputation for stock market crashes. In this timeframe (since 1996) the worst month for both the international indices I track has actually been August and the best April.

Update on Rates of Return

I just realized that I'd been computing the rate of return over the last 10 years incorrectly in earlier versions of this chart that I'd posted. The new calculation increases the variance a bit - so low and high returns are lower and higher than before. The graph only goes to the end of December. The rate of return over the last ten years was almost 10%. So far this month we are now over 10%. But these returns from the end of the tech crash market slump in 2002. Going forward, returns will fall again so that 6% seems like a more realistic long-term RoR. At the moment 2-3% is a realistic return for the S&P 500. Yes, that includes dividends. Of course, at some point this decadal bear market may end and a new bull market become apparent and RoR increase.

Facebook Investor Peter Thiel Makes Thousand-fold Return

The fall in Facebook shares since the IPO isn't much of a worry for the earlier investors. Peter Thiel sells 80% of his Facebook shares for $400 million. He bought his stake in the company for $1/2 million. Another way to look at it is that he made a 137% p.a. return in the eight years since investing.

P.S. 22 August

Actually, he made more than a 2000-fold gain.

Bill Gross is Wrong

In this piece, Gross says that stocks have had a 6.6% real return historically and asks how that can continue if the GDP grows at only 3.5% per year.

The answer is simple - you get a dividend of 3% and the value of your stock goes up by 3.5% in real terms to reflect the growth of future dividends in line with the growth of the economy. If you don't pay any taxes and reinvest all your dividends, the value of your asset grows at 6.5% per year and you would own an increasing share of the stock market. But in the long-run no-one can do this. At least they haven't. Even endowments like Harvard spend some of their earnings all the time. The stock prices of companies that don't pay dividends but make normal profits would go up at 6.5% per year. Berkshire Hathaway is a company that hasn't paid dividends for more than 40 years and its stock price has gone up enormously. At some point the model of buying more and more companies will run out of steam. In fact, I expect that after Buffett dies the managers will end up breaking up the company.

Lost Decade Update

Back in September last year I discussed the idea that the last decade was a lost decade as far as investment returns were concerned. Back then the MSCI World Index was outperforming the S&P 500 with almost twice the rate of return over the last ten years. And I had been outperforming both indices over that timespan. Now, 10 months on, the picture has changed a lot thanks to the ongoing European crisis and the underperformance of the Australian market:



The S&P 500 has caught up a lot and now has a 4.42% rate of return over the last decade, which at least beats inflation. This partly reflects, however, that June 2002 was in the depths of the "tech crash". The MSCI is now on 5.2% up from 4.15% last September. And I am at 4.54%, down from 4.61%.

Getting on Track for a Sustainable Retirement: A Reality Check on Savings and Work

A recent paper by Wade Pfau really overturns some standard ideas about retirement planning. The usual story includes picking a "number", planning how much you need to save to get to the number and starting early with saving in order to benefit from compounding interest. This table from the paper:



shows what percentage of final retirement assets could be explained by accumulated wealth a given number of years before retirement for a bond-stock portfolio simulated over the last century or so for the US. Even with an all bond portfolio only 43% of final wealth could be explained by accumulated assets 10 years from retirement. For stock oriented portfolios very little of the variation could be explained. It's all down to the luck of the market returns in the final 10 years. So tracking net worth doesn't really help much in telling you how much you'll have to retire on after all... Early compounding doesn't make much difference because there isn't much wealth to benefit from compounding.

So what does Pfau recommend? Calculating a minimum safe savings rate based on age, accumulated assets, and allocation. For someone of 55 years old who has saved 4 times their salary and wants to replace 50% of their salary and retire in 10 years and has 60% in stocks, the minimum safe savings rate is 52% of income! For a 50% chance of success of achieving a sustainable retirement only an 18.2% savings rate is needed. Having more wealth earlier does help reduce these rates. Not because of compounding but just in terms of piling up more savings. There are more analyses in the paper of the effect of retiring later etc.

Was There a Lost Decade?

An interesting article in the NY Times about whether investors should have had a lost decade - losing money over the last ten years or for the first decade of the 21st Century. The author argues that if you were properly diversified then you wouldn't have lost money. So I decided to check it out based on the data I have collected. I have been keeping accounts and returns on investment in a spreadsheet since 1996, so I have several years of data shown in the graph above. As at the end of last year the S&P 500 had a very meagre positive return over the previous decade. It's a little better now because it is coming off a lower base in 2001 in the tech crash. After inflation you would have lost money investing in the index. The MSCI World Index though had a return of 3.07% and would have about broken even after inflation. I was a little ahead with 4.31%. Right now I'm at 4.57%, MSCI at 3.77% and the S&P500 at 1.52%. So the author's point about diversification holds up. Back in the depths of the GFC all three were losing money. If you had a lot of relatively safe bonds you might have been above water. Not if you had a lot of Australian Dollars which fell dramatically in value in the crisis. So it depends what period you pick exactly, what result you get.

Are Tech Stocks Cheap?

Most large cap tech stocks now have fairly low price earnings ratios and some are even paying big dividends:



Historically, tech stocks paid very low or no dividends and utilities great big fat ones. From Tickersense.

Sovereign Credit Ratings

With all the talk about the U.S. credit rating possibly being lowered to AA from AAA due to the debt ceiling debate debacle, I was wondering what the credit ratings of other countries were as a point of comparison. Wikipedia, helpfully has a list. According to Standard and Poors, other AA countries include:

Belgium
Bermuda
China
Japan
Kuwait
New Zealand
Qatar
Saudi Arabia
Slovenia
Spain
Taiwan

Not such a bad club to be in. Some are countries with un-democratic governments but low debt like China and others democracies with very high debt levels like Japan. Apart from the U.S., the AAA club currently includes:

Australia
Austria
Canada
Denmark
Finland
France
Germany
Guernsey
Hong Kong
Isle of Man
Liechtenstein
Luxembourg
Netherlands
Norway
Singapore
Sweden
Switzerland
UK

These are democracies with reasonable debt levels and good track records or tax havens. Yes, Hong Kong is a tax haven. I don't think that Singapore counts as a tax haven and isn't a democracy really (despite having elections) but it does seem to have very good and reliable financial regulation.

Exchange Traded Actively Managed Funds - the Big Deal is that Foreign Investors Can Buy Them

This Wall Street Journal article goes on about how an exchange traded managed fund like the proposed PIMCO Total Return ETF is not big deal. It is, however, a big deal to foreign investors. Only US residents can buy units in unlisted US mutual funds. But anyone can buy US stocks in the secondary market. It would be nice if there were more of these.

Buffett Advises Against Long-Term US Bonds

Buffett speaking in India

In a recent post I referenced the Credit Suisse report that argues quite reasonably that long-term bonds are unlikely to be a good investment going forward. Of course, if you believe in efficient markets the prices of long-term bonds should already reflect that interest rates will rise in the future and, therefore, buying long-term bonds now should still be an OK investment. US short-term government bond interest rates remain near zero, but interest rates on 30 year bonds have risen significantly since the depths of the financial crisis:



History suggests though that the adjustment is insufficient. The Credit Suisse Report shows that there are long periods where bonds do not beat inflation in most countries with the partial exception of Switzerland. Now Warren Buffett warns against owning long-term US government bonds. His concern is both that inflation will reduce the real value of the bonds and that the dollar will fall in value against other currencies due to inflation in the US. It's hard to imagine the US Dollar falling in value a lot against the Euro, Pound, or Australian Dollar given how cheap things are in America but against developing country currencies such as the RMB that is possible.

My mother has a short-term USD bond fund and a longer-term Sterling related bond fund. We do want to reduce both of these and especially the latter. Snork Maiden and I have a variety of exposures to bonds though the total is only a small apart of our portfolio. The exposure is only 5.7% of net worth in total. The most significant types of bonds are Australian fixed interest and US corporate and mortgage related bonds. The latter are the main holdings of the CREF bond market fund, which did surprisingly well through the financial crisis (we should have had more of it):



This small level of exposure should be safe I think and I don't intend to lower it.