The holdings that did not change from the previous quarter were as follows:
American Express Co (AXP) was over 151.6 million shares
Bank of New York Mellon Corp (BK) was 1,793,915 shares
Comdisco Holdings (CDCO.PK) was roughly 1.5 million shares
Coca Cola Co. (KO) was right at 200 million shares
Costco Wholesale (COST) was 4,333,363 shares
ConocoPhillips (COP) is roughly 29.1 million shares
Gannett Co. (GCI) was 1.74 million shares
Dollar General Corporation (DG) was 4.497 million shares
General Electric Corp. (GE) was 7.777 million shares
GlaxoSmithKline (GSK) was 1.51 million shares
Ingersoll-Rand (IR) was 636,600; same as last quarter, but still way down from prior reports.
M&T Bank Corporation (MTB) was 5.38+ million shares
Mastercard Inc. (MA) was 405,000 shares
Moody's Corp. (MCO) was 28.4 million shares; same as last quarter, but way down from prior quarters and years.
Procter & Gamble (PG) was 76.7 million shares
Sanofi-Aventis (SNY) was 4.06 million shares
Torchmark Corp. (TMK) was more than 4.2 million shares
US Bancorp (USB) was roughly 69 million shares
USG Corp. (USG) was 17.072 million shares
United Parcel Service (UPS) was 1.429 million shares
Wal-Mart Stores Inc. (WMT) was just over 39 million shares
Washington Post (WPO) was over 1.72 million shares
The holdings with larger positions were as follows:
CVS Caremark Corporation (CVS) was a LARGER POSITION at 7.106 million shares
DirecTV (DTV) was a LARGER POSITION at 20.348 million shares
General Dynamics Corporation (GD) was a LARGER POSITION at 3.877 million shares
Intel Corporation (INTC) was a LARGER POSITION at 11.495 million shares
Visa Inc. (V) was a LARGER POSITION at 2.865 million shares
Wells Fargo & Co. (WFC) was a LARGER POSITION AGAIN at more than 383 million shares
The new holdings were as follows:
Da Vita Inc. (DVA) was a NEW POSITION at 2.684 million shares
International Business Machines Corporation (IBM) was 63.905 million shares
Liberty Media Corporation (LMCA) was a NEW POSITION at 1.7 million shares
The holdings with lower positions were as follows:
Johnson & Johnson (JNJ) was a LOWER POSITION at about 29 million shares; way under the peak of 62 million shares at one point in prior quarters
Kraft Foods (KFT) was a LOWER POSITION at about 87 million shares
Verisk Analytics, Inc. (VRSK) was a LOWER POSITION at 3.445 million shares
The holdings that were eliminated from the portfolio were:
Exxon Mobil Corp. (XOM) was ELIMINATED FROM THE PORTFOLIO
RISK RETURN MAP
Berkshire's portfolio without the new holdings listed above on a volatility-return map shows a historical average return of 17.3% return and 17.2% volatility (the yellow dot). Note that CVS Caremark Corp, Comdisco Holdings, Dollar General Corp, and Verisk Analytics have been excluded from the optimization calculations:
Click to enlarge
Berkshire has added three new stocks. Do these provide efficiency advantage to the overall portfolio? The Light Green, Peach, and Magenta square dots in the figure below represent the Berkshire portfolio when it is adjusted by adding increments of 10% in Da Vita, IBM, or Liberty Media (respectively), while proportionately reducing the original mix by the same percentage.
On average, adding Da Vita or Liberty Media increases volatility, but also increases average return. Adding IBM, however, increases volatility, while decreasing average return. It would be interesting to see if Berkshire reduces its IBM holdings in due course.
Now, let us view Berkshire's portfolio with the new holdings included. The yellow dot below shows the portfolio to have an average return of 17% and volatility of 17%. As discussed above and as seen in the ratio of return over volatility, the addition of the three new holdings do not add efficiency advantage to the portfolio.
Berkshire's portfolio as it stands on Feb 23, 2012, is shown in the volatility-return map below. The gap between the portfolio and an efficient portfolio with the same volatility has increased from 8.4% to 10%.
That said, the Berkshire portfolio remains very close to a portfolio with simple diversification, i.e. equal weights in each of the 30 stocks. The equal weighted portfolio (shown below) has a historical average return of 19.2% with a volatility of 18.5%.
Invest like Buffett?
Since putting equal weights in each of the 30 stocks above yields a volatility and return that is very close to the Berkshire volatility and return, is simple diversification all you need?
Let's see if the unconstrained optimized portfolio beats the equal-weights portfolio going "back into the future", i.e. we go backwards in time, do an optimization, and see how things work out going forward from that time.
In the two graphs below, we exclude Visa and Liberty Mutual in order to lengthen the amount of historical data that goes into the calculation. We use 100 as our starting portfolio amount on 31 March 2008 for easy interpretation.
It can be seen that the unconstrained optimized portfolio (denoted by the blue bars) beats the equal-weights portfolio (denoted by the red bars) every year for the next three years. At the end of March 2009, the blue bar was down 26%, but was nevertheless 19.3% more in value than the red bar. At the end of March 2010, it was up 39.4% and by end-March 2011, the blue bar was up 7.5% from a year earlier, 3.1% more than the red bar.
In the graph below, the starting point is a month later on 30 April 2008. The unconstrained optimized portfolio again beats the equal-weights portfolio every year for the next three years.
And the list goes on. But there is a reason why the optimized portfolio works so beautifully on Berkshire's list of stocks - it has to do with the concepts of "normality and "return to the mean", which I will discuss in my next article.