A couple days ago someone who follows my Twitter feed asked me what criteria I had used to pick the 16 JNets I talked about in a recent post. He referenced that there were “300+” Japanese companies trading below their net current asset value. A recent post by Nate Tobik over at Oddball Stocks suggests that there are presently 448 such firms, definitely within the boundaries of the “300+” comment.
To be honest, I have no idea how many there are currently, nor when I made my investments. The reason is that I am not a professional investor with access to institution-grade screening tools like Bloomberg or CapitalIQ. Because of this, my investment process in general, but specifically with regards to foreign equities like JNets, relies especially on two principles:
- Making do with “making do”; doing the best I can with the limited resources I have within the confines of the time and personal expertise I have available
- “Cheap enough”; making a commitment to buy something when it is deemed to be cheap enough to be worthy of consideration, not holding out until I’ve examined every potential opportunity in the entire universe or local miniverse of investing
That’s kind of the 32,000-ft view of how I arrived at my 16 JNets. But it’s a good question and it deserves a specific answer, as well, for the questioner’s sake and for my own sake in keeping myself honest, come what may. So, here’s a little bit more about how I made the decision to add these 16 companies to my portfolio.
The first pass
The 16 companies I invested in came from a spreadsheet of 49 companies I gathered data on. Those 49 companies came from two places.
The first place, representing a majority of the companies that ultimately made it to my spreadsheet of 49, was a list of 100 JNets that came from a Bloomberg screen that someone else shared with Nate Tobik. To this list Nate added five columns, to which each company was assigned a “1” for yes or a “0” for no, with category headings covering whether the company showed a net profit in each of the last ten years, whether the company showed positive EBIT in each of the last ten years, whether the company had debt, whether the company paid a dividend and whether the company had bought back shares over the last ten years. Those columns were summed and anything which received a “4” or “5” cumulative score made it onto my master spreadsheet for further investigation.
The second place I gathered ideas from were the blogs of other value investors such as Geoff Gannon and Gurpreet Narang (Neat Value). I just grabbed everything I found and threw it on my list. I figured, if it was good enough for these investors it was worth closer examination for me, too.
The second pass
Once I had my companies, I started building my spreadsheet. First, I listed each company along with its stock symbol in Japan (where securities are quoted by 4-digit numerical codes). Then, I added basic data about the shares, such as shares outstanding, share price, average volume (important for position-sizing later on), market capitalization, current dividend yield.
After this, I listed important balance sheet data: cash (calculated as cash + ST investments), receivables inventory, other current assets, total current assets, LT debt and total liabilities and then the NCAV and net cash position for each company. Following this were three balance sheet price ratios, Market Cap/NCAV, Market Cap/Net Cash and Market Cap/Cash… the lower the ratio, the better. While Market Cap/Net Cash is a more conservative valuation than Market Cap/NCAV, Market Cap/Cash is less conservative but was useful for evaluating companies which were debt free and had profitable operations– some companies with uneven operating outlooks are best valued on a liquidation basis (NCAV, Net Cash) but a company that represents an average operating performance is more properly considered cheap against a metric like the percent of the market cap composing it’s balance sheet cash, assuming it is debt free.
I also constructed some income metric columns, but before I could do this, I created two new tabs, “Net Inc” and “EBIT”, and copied the symbols and names from the previous tab over and then recorded the annual net income and EBIT for each company for the previous ten years. This data all came from MSN Money, like the rest of the data I had collected up to that point.
Then I carried this info back to my original “Summary” tab via formulas to calculate the columns for 10yr average annual EBIT, previous year EBIT, Enterprise Value (EV), EV/EBIT (10yr annual average) and EV/EBIT (previous year), as well as the earnings yield (10yr annual average net income divided by market cap) and the previous 5 years annual average as well to try to capture whether the business had dramatically changed since the global recession.
The final step was to go through my list thusly assembled and color code each company according to the legend of green for a cash bargain, blue for a net cash bargain and orange for an NCAV bargain (strictly defined as a company trading for 66% of NCAV or less; anything 67% or higher would not get color-coded).
I was trying to create a quick, visually obvious pattern for recognizing the cheapest of the cheap, understanding that my time is valuable and I could always go dig into each non-color coded name individually looking for other bargains as necessary.
The result, and psychological bias rears it’s ugly head
Looking over my spreadsheet, about 2/3rds of the list were color-coded in this way with the remaining third left white. The white entries are not necessarily not cheap or not companies trading below their NCAV– they were just not the cheapest of the cheap according to three strict criteria I used.
After reviewing the results, my desire was to purchase all of the net cash stocks (there were only a handful), all of the NCAVs and then as many of the cash bargains as possible. You see, this was where one of the first hurdles came in– how much of my portfolio I wanted to devote to this strategy of buying JNets. I ultimately settled upon 20-25% of my portfolio, however, that wasn’t the end of it.
Currently, I have accounts at several brokerages but I use Fidelity for a majority of my trading. Fidelity has good access to Japanese equity markets and will even let you trade electronically. For electronic trades, the commission is Y3,000, whereas a broker-assisted trade is Y8,000. I wanted to try to control the size of my trading costs relative to my positions by placing a strict limit of no more than 2% of the total position value as the ceiling for commissions. Ideally, I wanted to pay closer to 1%, if possible. The other consideration was lot-sizes. The Japanese equity markets have different rules than the US in terms of lot-sizes– at each price range category there is a minimum lot size and these lots are usually in increments of 100, 1000, etc.
After doing the math I decided I’d want to have 15-20 different positions in my portfolio. Ideally, I would’ve liked to own a lot more, maybe even all of them similar to the thinking behind Nate Tobik’s recent post on Japanese equities over at Oddball Stocks. But I didn’t have the capital for that so I had to come up with some criteria, once I had decided on position-sizing and total number of positions, for choosing the lucky few.
This is where my own psychological bias started playing a role. You see, I wanted to just “buy cheap”– get all the net cash bargains, then all the NCAVs, then some of the cash bargains. But I let my earnings yield numbers (calculated for the benefit of making decisions about some of the cash bargain stocks) influence my thinking on the net cash and NCAV stocks. And then I peeked at the EBIT and net income tables and got frightened by the fact that some of these companies had a loss year or two, or had declining earnings pictures.
I started second-guessing some of the choices of the color-coded bargain system. I began doing a mish-mash of seeking “cheap” plus “perceived quality.” In other words, I may have made a mistake by letting heuristics get in the way of passion-less rules. According to some research spelled out in an outstanding whitepaper by Toby Carlisle, the author of Greenbackd.com, trying to “second guess the model” like this could be a mistake.
Ultimately, this “Jekyll and Hyde” selection process led to my current portfolio of 16 JNets. Earlier in this post I suggested that one of my principles for inclusion was that the thing be “cheap enough”. Whether I strictly followed the output of my bargain model, or tried to eyeball quality for any individual pick, every one of these companies I think meets the general test of “cheap enough” to buy for a diversified basket of similar-class companies because all are trading at substantial discounts to their “fair” value or value to a private buyer of the entire company. What’s more, while some of these companies may be facing declining earnings prospects, at least as of right now every one of these companies are currently profitable on an operational and net basis, and almost all are debt free (with the few that have debt finding themselves in a position where the debt is a de minimis value and/or covered by cash on the balance sheet). I believe that significantly limits my risk of suffering a catastrophic loss in any one of these names, but especially in the portfolio as a whole, at least on a Yen-denominated basis.
Of course, my currency risk remains and currently I have not landed on a strategy for hedging it in a cost-effective and easy-to-use way.
I suppose the only concern I have at this point is whether my portfolio is “cheap enough” to earn me outsized returns over time. I wonder about my queasiness when looking at the uneven or declining earnings prospects of some of these companies and the way I let it influence my decision-making process and second-guess what should otherwise be a reliable model for picking a basket of companies that are likely to produce above-average returns over time. I question whether I might have eliminated one useful advantage (buying stuff that is just out and out cheap) by trying to add personal genius to it in thinking I could take in the “whole picture” better than my simple screen and thereby come up with an improved handicapping for some of my companies.
Considering that I don’t know Japanese and don’t know much about these companies outside of the statistical data I collected and an inquiry into the industry they operate in (which may be somewhat meaningless anyway in the mega-conglomerated, mega-diversified world of the Japanese corporate economy), it required great hubris, at a minimum, to think I even had cognizance of a “whole picture” on which to base an attempt at informed judgment.
But then, that’s the art of the leap of faith!