Please Note: Blog posts are not selected, edited or screened by Seeking Alpha editors.

2015 Performance

|Includes: Berkshire Hathaway B (BRK.B), INTC, LUK-OLD, MSFT

The portfolio returns 9.4% in 4Q/2015, versus S&P 500 return of 7.0%. Bringing the 2015 return to 4.6%, versus S&P of 1.4%. The result is subpar. Since this is my letter on Seeking alpha, I think it would be rational for me to discuss our allocation strategy. And then I would move on to what the fund current holds and why.

Our Allocation Strategy:

We are principally a long only value fund. Our positions are concentrated: usually no more than six to ten ideas. We constantly compare the possibility of return and loss to determine the size of an investment position. Most stocks we owned are ones that we follow for a long time. Generally, we follow three rules in our strategy:

· Zero possibility of permanent loss of capital.

· A business that we can understand and predict.

· A five-year accumulative owner's earning greater than at least 60% of company's current market capitalization.

The first two rules are simple for investors to understand. The third rule is what I think is as an appropriate way to value an investment in the near zero interest rate environment. By taking out the discount rate, we value a stock from an owner's perspective rather then worry about how Fed is going to with the rate. And we than simply compare the opportunity cost of owning one idea than the other. I simply cannot agree with most DCF model, where a slight change of discount or terminal rate can dramatically change the value of the businesses.

Our holdings:

Currently, our two largest positions, excluding cash, are Berkshire Hathaway and Intel.

Berkshire Hathaway:

We initiate a position in Berkshire Hathaway in the 4Q of 2015 and continue to buy more in 2016 as the share price increasingly approaches to its buy-back price, a level where Buffett consider a steal. Our average cost is $131, around 129% of Berkshire's 2015/3Q book value. An underperforming stock portfolio (AXP and IBM) and operating challenges at subsidies (BNSF and retail) all contribute to Berkshire's stock decline in 2015. We however, are glad that market present us an idea with minimal risk of capital loss and almost certain increase in intrinsic value.

Many investors are already very familiar with why Buffett will buy back the stock at 120% of book value. (Book value plus insurance float minus insurance goodwill). We feel no need to explain more. However, I just want write a little more on Berkshire's utility business.

We believe that Berkshire's utility business will over time become Berkshire's largest source of income. Our thesis is that Berkshire found a way to be the lowest cost provider in a very large industry. Berkshire benefit not only from its low borrowing cost, but also from its intelligent use of deferred tax liability. Prior to financial crisis, most of Berkshire's deferred tax liability derived from its unrealized stock gains. Since then, Buffet invested heavily in utilities. The deferred tax liabilities swell from $25.6 billion in 2009 to $70.4 billion in 2014, with large CAPEX in utilities and acquisition of BNSF contribute most of the net increase.

in $millions

2009

2010

2011

2012

2013

2014

Total Deferred Tax Liabilities

$25,636

$44,564

$47,645

$53,667

$65,873

$70,368

Investments

$11,880

$13,376

$11,404

$16,075

$25,660

$26,633

PP&E

$8,135

$24,746

$28,414

$29,715

$32,409

$34,618

Greg Abel, who runs Berkshire's utilities business, is not only a talent operator but also a shred investor. We excitedly wait him to pull next acquisition trigger. In the meantime, some of Mid-American's largest capex in renewable energy is close to finish will likely provide Berkshire earning a boost.

Intel:

Many investors view intel as no-growth company. We take a different view. We believe that the non-PC businesses, specially the data centers and memory businesses, will grow significantly over the next five years and generate vast majority of the company's profit. Our cost basis for Intel is around $30.2.

Intel dominates the data center business with more than 95% market share. The company's acquisition of Altera further solidify to its dominance. As a reminder, this is a $16-billion-dollar business ($18B if include Altera) with an operating margin around 50%. Intel laid out a strategy to grow the data center business at 15% for the next five years. While there is a lot skepticism if the goal is achievable, even taking conservative view of 10% annual growth and operating margin of 48%, this segment will generate well north of $11 billion operating profits by 2020. We believe the data center business alone worth at least $100 billion.

Intel's memory business, IMSF, is a 50/50 joint venture with Micron. The business currently generates less than $3 billion revenue for Intel with nearly no profits. We believe the memory business will be much bigger as NAND (especially SSD) become a mainstream memory storage format for both consumer and enterprise. IMSF is now mass producing their 1st generation of 3D NAND with a leading cost structure. While the NAND is a commodity business, by integrating NAND into CPU package will significantly improves Intel profit margin. An integrate memory solution also helps Intel defend off competition from ARM server providers. 3DXP point, a new non-volatile memory architecture with speed close to DRAM, can potential be disruptive by replacing both DRAM and NAND. Since 3DXpoint products won't be out until 2017, we remain conservative in our projection.

Intel's PC business has suffered for years as people shift to smart phones and tablet. To make the matter worse, the disastrous Windows 8 put many consumer and enterprise purchase on the sideline. However, the talk about "death of PC" is not true in our view. Even given all the headwinds, PC install base base globally are still growing. As Steve jobs long pointed out, "PC will be like trucks". We believe PC still has an important role in people daily life, especially in businesses. With Microsoft roll-out Win10, enterprise upgrades in 2016, coupled with the innovation in Skylake (Intel's 6th generation chips), and more hardcore PC gamers, we believe Intel's operating income from PC business won't shrink as much as street expected.

The transition of Intel does not come without challenges: a continuing sluggish PC demand, successful entry of ARM server, losing manufacturing lead in foundry, or unable to break through in mobile businesses. While the stock is trading below of our estimated intrinsic value, these variables also make the outcome of Intel less easy to predict than Berkshire. As a result, we trimmed one third of the position during the stock run-up in late 2015.

Other position changes in 4th quarter:

Microsoft: We invested in Microsoft during the 4Q of 2015 at an average price of $42. We exit at a price o at $54 as we find better ideas. We continue to believe in the long term transformation of Microsoft under the leadership of Satya Nadella and will re-entry the stock when it is more attractive.

Leucadia: We initiate at position in Leucadia at $16.8 in the 4Q of 2015. The thesis is that the stock is trading well below its tangible book value. We sold it recently at break-even price as we find more attractive ideas in the financial area.

Some thoughts on Technology:

While many deep value investors stay away from technology stocks, we do not hesitate owning them as long as they are understandable. One main thesis we believe is that each technology shift will bigger than the past one. Currently, we think there are four big technology trend converging within the next five years that will fundamentally change the way how we live. The four trends are: public cloud, flash, 5G, and big data software.

Public Cloud: The biggest technology shift in our view is the public cloud replacing traditional IT workloads. Today only 9% of IT workloads are in the public cloud. This number will go up several times as everyone predicts. One importance thing about public cloud is that it will enable a wave of new business model that wouldn't be possible in the past. For example, a standard bench mark to distinguish good business and bad ones in value investing is "return on net tangible assets". With organizations able to move their workloads to the public cloud, they can generate significant higher return on net tangible asset with very lower capex investment. Some can even achieve infinity returns on net tangible asset. When technology works, it works amazing.

Flash(NAND): Flash replacing hard disk drive(HDD) will be another technology shift. Computing used to be "CPU constraint", but now it is "memory constraint" as innovation in memory space lacks speed again in CPU. (So called memory wall). Traditionally, because of the high cost ownership of flash, consumer and enterprise could not massively adopt flash products even they are much faster than HDD. The introduction of 3D NAND dramatically change cost curve. NAND can now not only shrink (Moore's Law), but also stack (3D layers). As a result, the ownership cost of flash has dramatically reduced and are project to less or equal to HDD before 2018. A wave of new all-flash data centers will be build that will eventually benefit many end users.

5G: The deployment of 5G is also gaining momentum. Different from 4G area, where Qualcomm dominates the industrial standards, the whole technology industry now realize they can not miss the trend. Companies like Intel, ARM, Ericsson are either working on their own standards or forming alliance. Currently, mobile broadband in US has a download speed less than 100 mbps. In 2020, the first deployment of 5G will enable more than 1Gpbs download speed, with possibly 20Gpbs by 2025. 5G also has very low latency rate. This dramatic increase in networks will benefit not only consumer, buy even more in the industrial area.

Big Data Software: Lastly, many open-sourced software ecosystems are forming rapidly to leverage the advance of computing hardware. The most notable ecosystem is about big data analytics. Data scientist today are finding many exciting ways to get insight of the explosive amount of data. Notable open-source software platform such as Apache Hadoop, Spark and Flink has gathered hundreds of thousands of engineers around the world. Unfortunately, I have no expertise in software. While I follow the development, I generally do not invest in this area.

With all four technology trend converging at the same time, we can envision a world that big data is available everywhere. It will make many of today's hot topics such as autonomous cars, visual reality and augmented reality possible. It will also disrupt traditional industry such as retail and banking in a big way. In short, the impact of technology on our daily life is just getting started.

Some thoughts on macro environment:

While we are not macro investors, we care deeply about the economy. And today's environment makes us increasingly think that it is a time to "protect capital". The biggest problem the world is facing is the its ever increasing debt level. Especially debt from China.

Two simple charts can easily illustrate what we believe is the largest credit bubble ever created:

China Banking:

In Trillion RMB

2006

2010

2015

Banking Assets

34.5

94.2

194.1

GDP

18.6

39.9

67.7

China Housing:

In million square Meter

2007

2010

2015

Inventory

<100

241

738

Under Construction

2300

4055

7357

From 2006 to 2015, China's bank asset grows 5.6X times to 194 trillion. In the mean time, the GDP only grow 3.6X times to 67.7 trillion. As a result, an increase amount of credit translates to a less and less amount of productivity. The root of debt problem, in my views, is the huge pipelines of China's housing projects. The current housing inventory in china is 738 million square meter by official estimate. Home under construction, however, is almost 10 times over: a staggering 7.36 billion square meters.

The debt backing the construction is enormous, easily more than 30 to 50 trillion by my calculation. And these debt is deeply penetrated into the banking system. They are not only held by banks and institutions, but also retail investors through wealth management products- a term popularized in China in the past 5 years. In fact, some Chinese banks today earn more fees issuing WM products than they earn on net interest income. The hard part to deal with debt is not only how big the debt is, but what to do with the debt when default happens. If a deleveraging eventually comes, soured debt that are held among retail investors are much harder to buy-back compared to subprime debt that were able to be brought back by the Federal reserve.

In the meantime, the country is compounding their mistakes by many failed and conflicting policies. The most significant one is defending the Yuan after letting it to devalue in the first place. With PBoC draining $1 trillion from its forex so far to defend a mass devaluation, the central bank is essentially shrinking its balance sheet and doing a tightening. On the other hand, domestic activities still show dramatic increase in credit expansion. We believed that China simply can not have two conflicting monetary policies at the same time: maintain stable currency while growing credit. While we cannot forecast the eventual outcomes, the situation is severe and imminent. A correction in housing price in China can be catastrophic for the banking system. As Soros says, we are probably at best in a deflation scenario. We are very cautious with investments that has a large Chinese exposure.

Looking forward

While our individual stock picking method has less to do with the general macro economy, we do observe that the market is heading into a lot volatility. We do think this is a period to preserve capital. In the meantime, we are eager that if market can present us more buying opportunities. I look forward to you again on 2nd half of 2016.