Sunday, February 15, 2015

Net-nets aren't necessarily value investments

A net-net is a company whose market capitalization is less than two-thirds of its current assets less all liabilities. Ben Graham recommended net-nets as investments, and he and his illustrious disciple Warren Buffett have enjoyed a lot of success buying them. By now Graham and Buffett have both moved on from net-net investing, but it remains a popular strategy among up-and-coming value investors.

Last year, I wrote a skeptical post about statistically-cheap stocks. I quoted a message board commenter, "somrh," who wrote:

Part of what I'm interested in is whether or not value strategies really are 'value' strategies. Some of them seem to work more like short-term mean reversion strategies that may be really lousy over the long-haul.

I was curious to see if this is true of net-nets, so I informally backtested the idea using Value Investors Club. My backtest is unscientific because 1.) VIC doesn't provide a comprehensive survey of net-nets and 2.) I might have missed some of the net-nets that have been written up there. Nonetheless, I'm confident enough about the results to make a few conclusions:

• Many of the net-nets that give investors strong returns are acquired or liquidated within a year or two of becoming net-nets.

• Most net-nets that aren't quickly acquired or liquidated have bad long-term returns. These long-term losers may have good short-term performance, however.

• The performance of net-nets reflects the performance of the broad market to a greater degree than most investors appreciate.

• It may be possible to identify net-nets that will perform better than the net-net average.


Details of the backtest

I looked through VIC's archives for net-nets that were written up between January 1, 2000 and December 31, 2009 and found 55 of them. I divided the 55 net-nets into two groups based on their internal rates of return: 20 winners that had IRRs above 10%, and 35 losers that had IRRs below 10%.

10% may seem like an arbitrary boundary, but I found that the practical distinction was much greater: only a handful of stocks had IRRs near 10% . All but a few of the losers had negative IRRs, while most of the winners had IRRs above 15%.

For companies that were acquired or liquidated or went bankrupt, I measured the IRR from the date of the writeup until the company ceased to be public. For companies that are still public, I measured the IRR from the date of the writeup until today.

The backtest includes three stocks that technically weren't net-nets: two business development companies and a subprime auto lender. I felt comfortable including them because I thought their assets, while not classified as current for accounting purposes, could be liquidated much like inventory or receivables. For example, the auto lender's main assets were loans maturing within three years.

The backtest excludes one net-net that was a Chinese reverse-merger fraud, two net-nets that were involved in litigation that threatened to wipe out their assets, and three insurance companies that traded at less than 50% of book value.

These inclusions and exclusions don't significantly change the results.


Results of the backtest

The 20 winners were:
acls ascx avgn brnc cacs fc feim insp ksw lqid mckc mhh nick plus prls rhdgf scnya slnt ticc wilcf

The 35 losers were:
aab.to alvr aql.to awx czq.to dagm ebsc edci edvmf gigm gnom grvy gtsi hmb.v jlmc laco lens lgvn lxnt maxw mlnk mmco myrx msel mvc netp pme.to pmry snkty spor taa trnt txcc wnmla wqni

Five of the winners were acquired or liquidated within a year of being written up: ascx avgn lqid mckc slnt. brnc was acquired within two years of being written up.

By contrast only one loser, gnom, was acquired within a year. gnom was purchased at a significant premium to its market price, but the acquiror paid with stock that subsequently lost most of its value.

While the losers all had bad long-term returns, 19 of them had temporary returns of 100% or more from their writeup price: aab.to alvr awx czq.to dagm edvmf gigm grvy gtsi laco lgvn maxw mlnk msel pme.to pmry txcc wnmla wqni.

Almost all of the net-nets were written up during two periods: from 2000 to 2003 and in the twelve months following Lehman Brothers' September 2008 bankruptcy. Only a few net-nets were written up between 2004 and mid-2008. Only one net-net, rhdgf, was written up in 2007.


My opinions

Most net-nets are cheap for a reason: they have structural problems that prevent them from earning adequate returns on investment. Shareholders may profit if they're quickly acquired or liquidated, but the longer they stay in business, the worse shareholder returns are likely to be.

The flip side of this is that more than half of the VIC net-nets that did poorly long-term had strong short-term performance. As I suggested in last year's post, "Rather than measuring outperformance, [studies claiming that net-nets beat the market] may be measuring cheap stocks' ability to stage a dead cat bounce."

I don't think it's possible to divorce the net-nets' performance, at least their short-term performance, from the market environment. Almost all of the net-nets were written up in 2000-02 (a period when statistically cheap stocks outperformed the market), in 2003 (the beginning of an explosive rally), and in 2008-09 (the beginning of another explosive rally).

The one net-net that was written up in 2007, rhdgf, experienced a 60% drawdown after Lehman's bankruptcy. I take that as evidence-- admittedly, very thin anecdotal evidence-- that net-nets' strong historical returns are more a function of market environment than undervaluation. I'd be curious to learn how the net-nets that appeared in 1930 did over the subsequent two years-- my guess is their performance was awful.

Finally, I think it's possible to say ex ante that some net-nets are better investments than others. Five of the winners were profitable when they were written up and also had been profitable historically. Only three of the losers demonstrated consistent profitability, even though the losers were more numerous.

Most of the net-nets that were acquired or liquidated showed signs that an acquisition or liquidation was likely. ascx, avgn, gnom, and slnt hired advisors in order to pursue "strategic alternatives." ascx, avgn, gnom, and mckc implemented significant measures to curtail losses. Activist investors targeted avgn and lqid. I suspect that investor fatigue is what made it possible to buy these soon-to-liquidate companies significantly below liquidation value.

1 comment:

Note: Only a member of this blog may post a comment.