Since last week’s Japanese liquidation value: 1932 US redux post, I’ve been attempting to determine whether the historical performance of Japanese sub-liquidation value stocks matches the experience in the US, which has been outstanding since the strategy was first identified by Benjamin Graham in 1932. The risk to the Japanese net net experience is the perception (rightly or not) that the weakness of shareholder rights in Japan means that net current asset value stocks there are destined to continue to trade at a discount to net current asset value. As I mentioned yesterday, I’m a little chary of the “Japan has weak shareholder rights” narrative. I’d rather look at the data, but the data are a little wanting.
As we all know, the US net net experience has been very good. Research undertaken by Professor Henry Oppenheimer on Graham’s liquidation value strategy between 1970 and 1983, published in the paper Ben Graham’s Net Current Asset Values: A Performance Update, indicates that “[the] mean return from net current asset stocks for the 13-year period was 29.4% per year versus 11.5% per year for the NYSE-AMEX Index. One million dollars invested in the net current asset portfolio on December 31, 1970 would have increased to $25,497,300 by December 31, 1983.” That’s an outstanding return.
In The performance of Japanese common stocks in relation to their net current asset values, a 1993 paper by Bildersee, Cheh and Zutshi, the authors undertook research similar to Oppenheimer’s in Japan over the period 1975 and 1988. Their findings, described in another paper, indicate that the Japanese net net investor’s experience has not been as outstanding as the US investor’s:
In the first study outside of the USA, Bildersee, Cheh and Zutshi (1993)’s paper focuses on the Japanese market from 1975 to 1988. In order to maintain a sample large enough for cross-sectional analysis, Graham’s criterion was relaxed so that firms are required to merely have an NCAV/MV ratio greater than zero. They found the mean market-adjusted return of the aggregate portfolio is around 1 percent per month (13 percent per year).
As an astute reader noted last week “…the test period for [the Bildersee] study is not the best. It includes Japan’s best analog to America’s Roaring Twenties. The Nikkei peaked on 12/29/89, and never recovered:”
Many of the “assets” on public companies’ books at that time were real estate bubble-related. At the peak in 1989, the aggregate market price for all private real estate in the city of Tokyo was purportedly greater than that of the entire state of California. You can see how the sudden runup in real estate during the bubble could cause asset-heavy companies to outperform the market.
So a better crucible for Japanese NCAVs might be the deflationary period, say beginning 1/1/90, which is more analogous to the US in 1932.
To see how the strategy has performed more recently, I’ve taken the Japanese net net stocks identified in James Montier’s Graham’s net-nets: outdated or outstanding? article from September 2008 and tracked their performance from the data of the article to today. Before I plow into the results, I’d like to discuss my methodology and the various problems with it:
- It was not possible to track all of the stocks identified by Montier. Where I couldn’t find a closing price for a stock, I’ve excluded it from the results and marked the stock as “N/A”. I’ve had to exclude 18 of 84 stocks, which is a meaningful proportion. It’s possible that these stocks were either taken over or went bust, and so would have had an effect on the results not reflected in my results.
- The opening prices were not always available. In some instances I had to use the price on another date close to the opening date (i.e +/1 month).
Without further ado, here are the results of Montier’s Graham’s net-nets: outdated or outstanding? picks:
The 68 stocks tracked gained on average 0.5% between September 2008 and February 2010, which is a disappointing outcome. The results relative to the Japanese index are a little better. By way of comparison, the Nikkei 225 (roughly equivalent to the DJIA) fell from 12,834 to close yesterday at 10,057, a drop of 21.6%. Encouragingly, the net nets outperformed the N225 by a little over 21%.
The paucity of the data is a real problem for this study. I’ll update this post as I find more complete data or a more recent study.
Also, with regards to the study you quoted, the difference between buying net nets at a 1/3rd margin of safety vs. just below NCAV is huge. Returns are significantly different for the two groups… to the point where it’s almost not worth adopting the strategy at all.
Evan
LikeLike
Any strategy that beats an index by 21% is pretty impressive — though I’m wondering why you had such a short time frame. Any chance we’ll see a much longer study from Greenbackd?
LikeLike
This is actually pretty convincing. Through the worst bear market in 70 years, a mechanical strategy did almost as well as cash.
Or maybe better than cash: any dividends? Probably hard to get that info. But given the low return on cash over the period, any dividends would change the relative attractiveness quite a bit.
LikeLike
sorry i meant Beg-Sept 2008 price on that NOT end-Sept closing price
LikeLike
Thanks, Jamie. I’ll run it again.
LikeLike
Monsieur Greenback’d
Good work, thought I would help you fill in the gaps.
I pulled below off factset.
(BTW one cheap way to grab Japan stock data is
http://finance.yahoo.co.jp/
and drop in the stock code in the first bar, hit the yellow button and it gives you the last price – it’s in Japanese but it’s one of the best resources free on the web otherwise you’ll be forking out for BBerg or Factset subsciption)
OK, data organized as – Name/ticker; Sep-end 2008 closing price; feb 9th close;
6646 – Y168:Y168 (positive EV)
2423 – Y27,000:Y21,000 (still trading @ neg EV btw)
4642 – Y190:Y109 (still neg EV)
5921 – Y270:Y261 (positive EV)
9992 – Y290:Y205 (still neg EV)
5450 – Y250:Y150 (positive EV)
4352 – Y2285:Y2495 (this is called Nihon Industrial Hldgs – are you sure the ticker is right?) (positive EV)\
1734 – Y113:Y135 (negative EV)
6797 – Y510:Y385 (positive EV)
– Seki technotron appears to have been bought out bid price was Y155/share
– Sotec was bought by Onkyo Corp. (parent co.) via a share swap in July 2008 for %15.1M for the 48.6% stake it didn’t own
– Fujistu business systems was bought by Fujitsu Ltd. (parent co.) via a share swap in Aug 2009 for $280M for the 46.8% stake it didn’t own
– Kyoshin Technosonic merged with USC Corp. via a share swap in Oct 2009 for $40M deal
7877 – Y150:Y95 (positive EV)
– EC-one inc I have no data on – it was delisted in Feb 2009, according to its website
4625 – Y360:Y295 (positive EV)
9941 – Y214:Y116 (positive EV)
– Shinkigosei was bought by Blue River via a TOB in Oct 2008 for Y400/share for the 27.1% stake it didn’t own
So quite a few of the net-nets with a parent-child connection appear to have been bought out.
LikeLike