Monday, March 5, 2012

Apollo Investment vs. Berkshire Hathaway: A Response

A comment at Seeking Alpha to Why I Sold Apollo Investment - Part III seems to invite a whole article in response.

OJL13 wrote (and here I paraphrase) that at Berkshire's high price for Class B shares, small investors can afford few shares, which stand little chance of outperforming Apollo when Apollo actually pays a dividend.

When I first noticed Berkshire Hathaway in the early 1980s, it was trading at $3,850 a share and doing an outstanding job of making money. I was attracted, but thought then as OJL13 does now: how many could I expect to buy? What was worse, commissions in the early '80s were in the hundreds of dollars, and there were penalties for what they called "odd-lot" trading – that is, buying or selling a share count not divisible by a "round lot" of 100. So I didn't buy. Shortly after, a news story covered its hitting the $4,000 mark. But what's $150 out of $3,850, right? Not even enough to pay a commission. (At least, in the early '80s)

Fast forward a bit. Those once-$4k shares (now referred to as the Class A shares) have a book value per share of about $100k, and trade above book. A new Class B share – which danced with $4k itself a few years back before a 50:1 split – now allows investors to buy near $80 per slice. And today, that purchase is pretty cheap: buy/sell transactions can now be accomplished at a broker of your choice for under $10 each, at any share count. The odd-lot penalty is dead.

Now, the investor's real questions are:
  1. Does the investor require liquidity (currently spendable money) that you don't want to realize by selling a partial stake?
  2. Does the investor desire immediate taxation on dividends paid by a company losing share value that can result in a tax deduction only in the future, at some hypothetical and potentially distant point of exit? (And does the investor intend risking lowered returns as net assets per share dwindle?)
  3. Does the investor believe that Apollo's management can produce a return – net of its high fees – that beat those of Apollo's alternatives? And here one notes that Apollo's alternatives include Berkshire, which is making broad leveraged bets (using both long-term derivatives and interest-free "float") in favor of the U.S. economy with a back-office overhead of only 24 employees, and is not saddled with an expensive external manager.
The problem raised by the reader really comes down to an issue Berkshire discusses in its 2011 Annual report. On pages 99-100, Berkshire frankly discusses the decision to retain earnings and the alternative, to pay dividends. Berkshire's discussion is spot-on, and are a must-read for those weighing the merits of demanding dividends from investments.

Investors' choice is simple: will management will make more money with after-tax earnings, or will an investor make more money with what will be left of the after-tax earnings after the investor then pays taxes on receipt of a dividend? If management can make more money with the earnings, investors want management to retain the earnings to provide investors not only the benefit of carefully-selected management (why else would an investor buy?), but the benefit of avoiding avoidable double-taxation required upon payment/receipt of the dividends. Dividends make sense in circumstances such as when (a) a tax-pass-through structure makes double-taxation avoidable if some dividend-payment test is met (e.g., the 90% distribution requirement of RICs like Apollo Investment and REITs like AGNC and MTGE), or (b) management can't think of anything more productive to do with the earnings than you can.

I think that the comparison of AINV against lower-overhead leveraged debt investments such as AGNC and MTGE shows that AINV isn't the best place to put money. Over the last year, the market has agreed:
Losing over 40% of investors' share price (to about $7) is not a small price to pay for a dividend that's been cut to $0.20 per quarter (~11% of the remaining $7 share price). This performance was lambasted not long ago by Selena Maranjian, who singled out AINV among high-paying stocks whose equity decay rendered investment pound-foolish.

Unless there is some underlying reason AINV should be thought erroneously undervalued by the market, AINV is a sell. Period. Its expensive management is producing net losses for shareholders, with no remedy in sight. The only question is where to invest sale proceeds. So, let's have a look at the Berkshire Hathaway Class B shares over the last year:

As discussed in Why I Sold Apollo Investment, Part III, Berkshire turned in this blah-looking share price performance while beating the S&P 500 in book value per share growth and acquiring outstanding companies with terrifically growing – and in 2011, record-setting – operating earnings. Berkshire's price is an erroneous undervaluation of its true value. Apollo's assets per share and dividend have – as discussed in Part I of the series – declined awfully over the period. The fact that Apollo's management is charging a huge fee to produce subpar results explains why it's a sell, whereas Berkshire's outstanding returns coupled with its paradoxial price decline explains why it's a buy. Of course, maybe it's not a buy for you. A near-term investment horizon would, for example, make it hard to sit patiently waiting for the market to properly value derivatives whose true value may remain a puzzle until the last of them expire in the late 2020s.

In a tax-deferred account, I've suggested MTGE as a replacement for AINV as a leveraged debt investment. American Capital Mortgage Investment shares the tax-efficiency of a tax-pass-through structure, and in a tax-deferred account this allows reinvestment without the irritation of taxes nibbling at one's reinvesting annual returns. In a regular (taxable) account, some investors will find their tax burden (as low as 0%, depending how broke one is and what tax credits and deductions are available in a given year) to be lower than that of corporations (35%), providing a diluted version of the same benefit.

In a taxable account, Berkshire's post-tax results are hard to beat for long periods of time, and the company has some real basis for a conclusion it is undervalued. Maybe not as undervalued as ACAS (manager of AGNC and MTGE), but that's a different article. As discussed in the original article, the Berkshire recommendation is based on (a) free leverage, (b) equity exposure, and (c) mispriced assets (the derivatives, float, and tax liabilities are all systematically overstated by GAAP in comparison to their real present value as liabilities). It is also based on something seemingly utterly absent at Apollo, which is quality management.

No comments: