Comment by ScottBurson | Hacker News Reader

Comment by ScottBurson | Hacker News Reader

There's a lot of failure to appreciate a simple fact: "1 share" is not a consistent, well-defined unit. It's just some arbitrary fraction of the value of a company, that fraction varying not only between companies but also over time, because of splits, secondary offerings, and buybacks. This seems like an obvious, straightforward fact that everybody knows, and yet historically, there has been a tendency to ignore it and to think the price of "a share" means something. Companies whose share prices are above $50 or so have been viewed as "blue chip"; those below $10 have been viewed as "speculative". There is thus a psychological component to the perception of share prices that is not justified by the mathematics of the situation; and much of this is a matter of convention. Even today, I believe that to continue to be listed on the NYSE or even NASDAQ, a company must not let its share price fall much under $1. The only effect of AMZN trading at $1730 is to make it difficult for small traders to trade the stock, since exchanges don't deal in fractional shares (they could, I suppose, but they don't).

We see this error in the way that Charles Dow and Edward Jones defined their index back in 1896 [0]. Averaging prices of shares in different companies is mathematically meaningless. But the author of this piece, having commented on this error, goes on to make it again: For example, for fiscal year 2017, Costco had earnings per share of $6.08. Amazon had earnings per share of $6.15. Costco’s market value is $91 billion; Amazon’s is $844 billion. EPS numbers aren't directly comparable any more than share prices are directly comparable. COST is trading around $209, less than 1/8 AMZN. Instead of EPS, we should be talking about earnings per dollar of market cap, which would be independent of the size of a share. On that measure, COST is outperforming AMZN by a factor of more than 8. (Along with juxtaposing EPS numbers from two different companies, the author seems to be committing a second error by suggesting that earnings and market cap should be related; they are not, at least not in any simple way.)

Your idea of "earnings per dollar of market cap" is essentially the same as "earnings yield", or inverse of the P/E ratio, which is indeed much more meaningful than EPS

Earnings yield is a measure of how expensive a unit of earnings is, and can be compared across companies as a heuristic for "cheapness".

Another metric is free cash flow yield: free cash flow / market cap. Using operating free cash or operating profit like EBITDA gets you the cost of operating cash flow excluding non-operating stuff like interest, taxes and non cash depreciation expense

You can also use enterprise value as the denominator (market cap plus debt less cash) to capture the full picture of a company's capitalization

By buying groups of stocks based on these metrics, rather than just index funds, you can blend aspects of passive and active investing. This is called "smart beta" and is an interesting emerging strategy

A good primer is "the little book that beats the market". The strategy is to buy "cheap and good" stocks and just hold them. Cheap is defined by high EBIT / EV (lots of earning power per unit of capital) and good by high ROIC (the business can turn money into more money). So you filter by these two metrics, then buy the top x companies, buying one stock / month over a year to mitigate timing risk, and hold

I believe that you need to buy about 20-30 of these companies then hold each for about a year, to get long-term capital gains, then reevaluate. Greenblatt claims that statistically this will work.

The Magic Formula:

https://en.m.wikipedia.org/wiki/Magic_formula_investing

Has anyone tried it?

I haven't tried it though I find it intellectually appealing. Actually implementing it requires some config (do you want to include all industries, or exclude some like finance as greenblatt suggests? What range of market caps will you consider? US only or emerging markets? Do you short companies on the bottom of the list or do long only?)

I just haven't spent the time thinking through those things...

There's a good article on more current results here: https://seekingalpha.com/article/2944796-beating-the-little-...

Short version is that the alpha has shrunk considerably.

The Acquirer’s Multiple is another similar formula:

http://acquirersmultiple.com

One must be careful to avoid a "value trap". Occasionally, a company is inexpensive for a reason, and that reason can be hard to discern at a glance.

That is of course true. However, I've noticed many seem to justify any stock price being low as the market already pricing in all the information. However when you look at AAPL's valuation 1-2 years ago, shares went down to a low of around $90, which was only like 6-7 FCF or so (it's a little complicated cause you had to subtract apple's cash hoards but also multiply by some uncertainty factor < 1 that represented the probability that they would be able to bring the money back).

Meanwhile you have companies like MSFT, etc which while growing at a fast rate, are far, far more expensive on a FCF yield basis.

Ultimately - and this is pretty obvious but worth stating - you have to come up with your own valuation for a given company when making investment decisions.

That's a risk, but ROIC is supposed to somewhat counter that by ensuring the company is creating earnings from its invested capital (i.e. It's a "good" company). Can also only include large cap stocks and diversify to remove some of this risk

Regarding "Instead of EPS, we should be talking about earnings per dollar of market cap", isn't that what the P/E ratio measures? Although, strictly speaking you are suggesting the reciprocal - an E/P ratio.

I believe P/E ratios are far more widely used for comparing companies.

EPS isn't a completely useless measure - it can be useful to shareholders in the specific company to provide some anchor for judging dividend policy and for comparing performance over time for that particular company.

But I agree that EPS is a useless measure for comparing different companies.

We should never be talking about earnings per dollar of market cap (PE, as you've correctly noted) - the far superior metric is fcf per dollar of market cap. (Even FCF isn't a flawless metric, since it can be affected by one-offs, but those are generally easy to account for, whereas earnings being a pure accounting metric is vulnerable to all kinds of fuckery)

can you expand on that? Is FCF free cash flow?

Yes. Let me know if you want me to like actually expand on it (versus just wondering what FCF stood for)

how is FCF (normalized per market cap) a better metric? What kind of fuckery is P/E vulnerable to?

Sorry I’ve been swamped so i can’t actually do reaearch and give a full response so I’ll start with a tldr and hopefully come back and give you a real answer later:

earnings are an accounting measure and there’s some flexibility as to whether a given expense is classified as capex, opex, r&d, etc each with different earnings implementations.

In general I prefer cashflow to earnings for that reason.

The next point is cash flow vs free cash flow, fcf factors in capex, etc giving you an accirate idea of how much capital is free to distribute to shareholders (free meaning available).

As one example, look at tsla’s financials. Their gross margins / earnings paint a way better picture than their fcf, because they engage in all the accounting practices that I vaguely referenced above

thank you for taking the time to write this

I disagree. A share is worth a portion of the business. That is a defined unit. The share price is what people will pay for a share of the piece of company. Baked into the price is the underlying assets (real value) in addition to the fuzzy speculation piece.

The share prices can be analyzed by looking at the financial reports. True value can be assessed by looking at assets minus liability, debts, dividend payment, etc.

This value per share then is compared to the buying price. There is a lot of speculation built into the price now a days. An investor does not have to buy at this real value + speculation price, the investor can sit it out. People are willing to pay for a share of this company, a non-defined amount, it's the only way to continuously offer a share of the company for sale.

Likewise people sometimes make the mistake of comparing prices for different cryptocurrencies. In the case of cryptocurrencies one should look at the market cap, which is the current price multiplied by the amount of coins or tokens available.

What’s the value of the company if not assets + all future earnings?

The market clearing price.

I mean, that's the theoretical value that you should pay for it, but markets don't always exactly match theory.

What about debt/liability?

I guess you might be referring to Enterprise Value ( EV ).

https://www.investopedia.com/terms/e/enterprisevalue.asp

Ultimately, the value of a company is what someone is willing to pay for it. So you could argue the value is EV + Premium.

Keep in mind that there is a bazillion ways of valuing a company. People look at growth potential of the company, the growth potential of the company's industry, the competition, external risks, buyout potential, etc. It's part business, part mathemagic and part voodoo.

What's the value of a 'future earning'?

And that is why psycologists are winning Nobel's for economics these days. Tomorrow they will hopefully play a bigger role in engineering and tech.

Reply on news.ycombinator.com