Corner of Berkshire and Fairfax Fund

[rkbabang]

What would be a good way to do it would be a link that takes you to a site, where you put in your COBAFF username and the password: "COBAFF" or something (just to keep random people from the web from voting even if they find the page.  Then it takes you to a form, where you have 2 columns of 10 rows.  You have 100 points to spend.  You put in a stock ticker in column 1 (or a $ for cash) and the amount you wish to put in it in column 2 for up to 10 stocks. Then hit submit. If your points you assigned are greater or less than 100 it tells you to try again.

 

 

The site then compiles all of the data and lists the stocks and the percentages of points that they have received by everyone who's voted.

 

That way you could do something like this (not my real choices):

 

$	5
TSLA	29
BAC	2
AAPL	11
FRFHF	1
BBRY	52

 

 

 

[Parsad]

If a CBF portfolio was housed on Covestor, board members (and others) could invest in it.

 

Something like this: http://covestor.com/mco-investments/margin-of-safety

 

I thought about Covestor actually. If this works well, I'd be open to starting a model portfolio on covestor. Fees collected would be donated.

 

I have no problem with that.  I would only have a problem if someone was profiting extensively from it.  Cover your costs, and pick a non-profit of your choice.  Good idea Ross!  Cheers!

[Ross812]

What would be a good way to do it would be a link that takes you to a site, where you put in your COBAFF username and the password: "COBAFF" or something (just to keep random people from the web from voting even if they find the page.  Then it takes you to a form, where you have 2 columns of 10 rows.  You have 100 points to spend.  You put in a stock ticker in column 1 (or a $ for cash) and the amount you wish to put in it in column 2 for up to 10 stocks. Then hit submit. If your points you assigned are greater or less than 100 it tells you to try again.

 

 

The site then compiles all of the data and lists the stocks and the percentages of points that they have received by everyone who's voted.

 

That way you could do something like this (not my real choices):

 

$	5
TSLA	29
BAC	2
AAPL	11
FRFHF	1
BBRY	52

 

rkbabang,

 

I like your idea. There are various survey websites that we maybe able to set something up like that. Maybe even a google docs spreadsheet would work but it could be thrown off if someone was dishonest. My only hangup with an external survey is there is another step to the voting which may drive down participation. Looking at the results, are they along the lines of what you expected? I'm wondering if we may be complicating something that will work itself out.

 

Also as for allocation. I am thinking we either pick the top 10 stocks, or keep all the stocks that received at least 3.5% of the vote. Then divide the number of votes for each stock by the votes garnered by the sum of the selected stocks to get a % allocation. The 3.5% rule would allow the portfolio to be flexible and get very concentrated if there was a large consensus (min. 5 stocks) or very diversified if there was little consensus (max. 28 stocks). I'm guessing the number of stocks in the portfolio would float around 16 in a normal distribution.  Thoughts anyone?   

 

[constructive]

I think the fact that over 100 people voted in a single day, and expressed strong consensus of 20+ votes on some of the top picks, suggests that the voting format in this post will probably work. Not perfect, but good enough and extremely easy to use.

 

Although I'm not the biggest fan of SHLD, BAC and cash, I think it would make sense to introduce some weighting. Maybe a 15 stock portfolio with full positions for the top 5, three quarters positions for the next 5, and half positions for the next 5.

[SharperDingaan]

All one need do is snap-snot the investment ideas section at the end of each quarter; then compare names & the number of posts. To get the pulse of the board, simply filter for the new names, & the old names with above normal posting activity. To get market timing add a bull/bear sentiment tracker to every investment idea. Data mining 101.

 

As the initial step focus on a baseline portfolio that is fully automated & algorithmic. Essentially portfolio weightings, & market timing, driven purely by formula - based on board posting activity. First time out it probably isn't going to be great, but over successive generations this should improve rapidly (think kaizen & multi-generation cell 'phones). It is what you measure against, & the process is a close cousin to the algorithmic search generators behind HF trading. Anyone on this board (or off it) can then compare their performance to the baseline portfolio; we get a training tool, & the board gets comparative advantage & promotion. Every CFA student will be using the board to practice on, & it will ensure a ongoing future contribution from trained investment professionals.

 

It would also be nice to break the baseline portfolio into regional portfolios such as NA, Europe, Asia, etc.  Expands the cross pollination of the board, broadens perspectives, & promotes the board globally.

 

 

[glider3834]

great idea, there are some smart investors on this board & I will be interested to see how the preferences look once we have had a good number of votes cast. Cash & AIG were my two favs.  ;)

[SharperDingaan]

The obvious 2nd step is integration of the RM strategies that are also part of this board. To do it you need a PM making the decisions, & you would compare results against the baseline portfolio. The difference is the value add, & it is wholly attributable to both the PM's skill & the RM techniques available to him/her. The fairest & most practical way to do this is via an actual real portfolio, with the benefits going to a charity, & the portfolio position published monthly. If board members wish to replicate, all they need do is monitor the month end position, & buy/sell accordingly in their own account - no fiduciary issues, taxes, etc.

 

It would be nice to have regional portfolios as well, & each portfolio managed by a PM local to that region. It crowd sources the burden, uses local knowledge, & allows board members to mix/match across regions to improve diversification. It also builds board membership, & harnesses the competitive bias of the investment business (business school/newspaper portfolio competitions) to the greater benefit of all.

 

Obviously a very big stretch at the moment, but no different to the typical project. Do the initial step well, to prove the concept; & the rest of it will fall into place naturally.

 

Grab the cohones, & squeeze!

 

[Packer16]

One question is what are we trying to do with this portfolio?  If we are trying to see if the wisdom of the CoBF crowd is correct we may want to keep the portfolio small and concentrated in a few names (maybe weighted by votes).  This will allow any signal to be separated from the noise.  Which is the biggest issue in mutual fund performance measurement.  Initially I am skeptical that such a system is going to work well because it is going to pick the favorite value stocks versus using some sort of value metric to measure the sentiment of the idea as reflected in the price.  We will see.

 

Packer

[Crip1]

Two separate thoughts:

 

1) Ross, thank you for taking this on. I really like the concept and will be very interested to see the results. There have been numerous suggestions from other board members on how to take this forward; I would only add that since this is your “baby”, feel free to do what you want.

 

2) I am a little surprised that FFH is leading MKL and, to a lesser extent, LRE. On a few FFH threads there is call for FFH to be more like MKL and/or observations that FFH is slowly morphing itself into a company more closely resembling MKL (higher focus on underwriting, less leveraged, more non-insurance businesses, etc). This suggests that MKL is the more evolved of the two companies. If so, then why would FFH garner nearly three times the votes of MKL? Valuation maybe? This is not a criticism of any voting nor is it campaigning for MKL. Candidly, if given X about of dollars and told that I have to invest all of it in either FFH or MKL, I would struggle with the decision. I just find it somewhat unusual.

 

Looking forward to watching this for the months/years to come. Thanks again, Ross.

 

-Crip

 

[John Hjorth]

Two separate thoughts:

 

1) Ross, thank you for taking this on. I really like the concept and will be very interested to see the results. There have been numerous suggestions from other board members on how to take this forward; I would only add that since this is your “baby”, feel free to do what you want.

 

2) <Omitted, here, John>

 

Looking forward to watching this for the months/years to come. Thanks again, Ross.

 

-Crip

+1! This intiative by Ross is much appreciated! The project has to start at some level/platform. Rome wasn't built in one day. Please also take into consideration, that running this project for Ross takes time away from managing own positions.

[SharperDingaan]

We concur, this is not our baby .. & we look forward to seeing it evolve.

 

Have fun!

 

SD

[arbcon]

Possibly we should start a thread on the next company we can convert to a baby berkshire?

[Ross812]

Thanks for the encouraging comments. I'd love to turn this into a community project at some point if board members want to continue the idea into something bigger. There are many members on here that have much more experience at setting up funds and non-profit framework, finding investors, and managing funds. If this someday evolved into a charitable fund I would hand the reins off to a board member who was capable of maximizing the charity's mission statement.

 

I'm sure as the idea evolves new metrics for balancing and allocation will emerge. After a few quarters of trial, we can figure out a better way of collaborating between board members. Perhaps, take the poll once a quarter and establish 10-15 chosen stocks, then have a few select board members develop their own replicable algorithms to balance, allocate, and time positions. The 'managers' could run their sister funds based off the crowd sourced stock list and see if their algorithm can beat the crowd sourced allocation percentages. 

 

The data collected so far is pretty interesting in its own right. The cash allocation change from quarter to quarter can be used for market sentiment like that at AAII. Changes in allocation could also help signal when ideas are reaching fair value. I suspect that we will observe the in early out early phenomena that plagues value investors. Look at the top 15 stocks right now, they look very reminiscent of Fairholme. I had no idea there was such a consensus on Sandridge, and I suspected Altius and Fortress paper have a lot more votes than we are seeing.

 

We concur, this is not our baby .. & we look forward to seeing it evolve.

 

All I would ask is any data gathered, the ultimate direction the 'fund', and any techniques developed by the community to further this idea remain transparent to the group. There are so many ideas, countless hours of research, and advice shared between the members of this community that the ownership of any fund based on the board's ideas really does belong to each and every member. 

 

Regards,

 

-Ross

[rjstc]

Two separate thoughts:

 

1) Ross, thank you for taking this on. I really like the concept and will be very interested to see the results. There have been numerous suggestions from other board members on how to take this forward; I would only add that since this is your “baby”, feel free to do what you want.

 

2) <Omitted, here, John>

 

Looking forward to watching this for the months/years to come. Thanks again, Ross.

 

-Crip

+1! This intiative by Ross is much appreciated! The project has to start at some level/platform. Rome wasn't built in one day. Please also take into consideration, that running this project for Ross takes time away from managing own positions.

 

+1 I always find Ross to be very transparent, helpful, & interesting. This is a great idea.

[Ross812]

With 164 votes in so far here is what the model portfolio would look like:

 

BAC  13.7%

SHLD 8.5%

AIG 8.3%

BRK.B 7.9%

FFH.TO 7.5%

SD 6.5%

FIATY  5.0%

LUK 4.8%

WFC 4.8%

AAPL 4.4%

LRE.L 4.4%

DTV 4.2%

OIBR 4.0%

MKL 3.1%

___________

Cash  12.9%

 

 

 

To get the allocation I took the number of votes the company received divided by the sum of votes garnered by the top 15 picks. I'd like to keep the poll open for a while longer. Hopefully we can get the voting up to 200 members.

[Ross812]

The fund went live last night with 172 member votes. Data collected from the poll can be viewed on the 'Allocation by Quarter' tab. The final allocation is as follows:

 

 

Link to portfolio:

 

https://docs.google.com/spreadsheet/ccc?key=0AivVdWOTQE2JdGQzYkFSQUg2eUR3UVRTcThFYWpBelE#gid=21

[racemize]

This is very cool.  Please keep going in the coming quarters/years!

[ASTA]

Cool nice job Ross812 going to be interesting how this develops.

[petec]

This is very exciting, but...what are we actually capturing when we filter by # of board posts?  Not the best ideas IMHO, but the ones that generate the most controversy and/or are the most complex.  No surprise to see banks, tech stocks, insurers on the list!

 

Clearly complexity and uncertainty can create great investment opportunities...but so can boredom and apathy.  Can anyone think of a way of generating ideas from the posts here that *don't* attract a lot of attention?

[SharperDingaan]

Looks great.

 

Can we add in the reference 1yr US T-Bill rate, & the reference levels for the exchange indexes these stocks trade on ? The ask is because as per MPT,  the portfolio should do better than the investable risk free rate (1 yr US T-Bill). And hopefully ....  it will also do better than the lowest cost practical alternative of simply buying the indexes in equal weightings.

 

SD

[matjone]

This is very exciting, but...what are we actually capturing when we filter by # of board posts?  Not the best ideas IMHO, but the ones that generate the most controversy and/or are the most complex.  No surprise to see banks, tech stocks, insurers on the list!

 

Clearly complexity and uncertainty can create great investment opportunities...but so can boredom and apathy.  Can anyone think of a way of generating ideas from the posts here that *don't* attract a lot of attention?

 

you could just track a portfolio of the least discussed investments on here. 

[constructive]

Awesome. I like the setup.

 

Here are some stocks that might make the ballot next time:

FNMAS (or other Fannie / Freddie preferred)

TPRE (or TPOU.L)

GS

GNW

LUKOY

DJCO

MIL

[Ross812]

Looks great.

 

Can we add in the reference 1yr US T-Bill rate, & the reference levels for the exchange indexes these stocks trade on ? The ask is because as per MPT,  the portfolio should do better than the investable risk free rate (1 yr US T-Bill). And hopefully ....  it will also do better than the lowest cost practical alternative of simply buying the indexes in equal weightings.

 

SD

 

Added the S&P 500 as a benchmark through SPY. $1M of SPY was bought at 164.6. All dividends will be reinvested.  I also added the 1-yr US T-Bill Rate on the 'Allocation by Quarter' tab. I added on starting prices for the TSX and LSE.

 

Awesome. I like the setup.

 

Here are some stocks that might make the ballot next time:

FNMAS (or other Fannie / Freddie preferred)

TPRE (or TPOU.L)

GS

GNW

LUKOY

DJCO

MIL

 

They will be added next time (Q4). The companies receiving no interest will be dropped from the poll and newly discussed ideas added.

[james22]

Might we add a consensus holding price as does dataroma (http://www.dataroma.com/m/g/portfolio.php)?

 

"Hold Price" is the term we use to indicate the price at which our Super Investors are comfortable owning a security. It is the price of the security on the portfolio date (usually quarter end). It's a reasonable indicator of a stock's undervaluation at that price, otherwise it would have been sold. We only provide this value when a security is owned by at least two Super Investors. In the case of reporting dates being different, we calculate a weighted average price.

 

Or maybe P/B preferable?

[mikazo]

Would the unrealized gains/losses on this spreadsheet be accurate? I think so, if it's pulling latest quotes from Google Finance. Interesting to see where this fund ended up, especially compared to SPY.