Far from it. As I've written before, celebrity branding makes me very nervous. The Ben Graham ETN is not a true total market fund. It takes huge risks and bets against a true market portfolio, with a non-transparent methodology that any investor would question.

But the Ben Graham ETN is almost beside the point.  The point is that ETFs give investors the means to express their precise views on the market.

My point would have better been made by examining all the ETFs in a given sector. Take, for example, the returns of the various healthcare ETFs in January.

The swing on YTD results is 8.24%, ranging from the iShares DJ US Medical Devices ETF (NYSE Arca: IHI), which was up 3.40%, to the PowerShares Pharma ETF (NYSE Arca: PJP), which was down 4.84%. 

If you look at the gradation of returns by industry segment, you can break it down more or less into industry categories:

• Medical Devices
• Biotech
• Broad-based Healthcare
• Pharmaceuticals
• Healthcare Services

I'm Monday morning quarterbacking here, but as a former biotech analyst, it's easy to explain these returns. 

If you examine which parts of the healthcare sector are most exposed to both the current economic downturn and general cost pressures, it would break out the same.  Both biotech drugs and medical devices are cost-insensitive: they are generally fixed-price products with limited or no generic competition, and somewhat limited replacement alternatives within their space (particularly on the biotech side).  Pharmaceuticals, on the other, are facing enormous generic competition with the ramp-up of programs like the $4 generics Wal-Mart plan.  Health care services, to round out the group, is more exposed to employment trends and broader cost-cutting measures at the corporate level.

That's not to say that these returns will follow a similar pattern in the future. But you could make an argument for one or another piece of the healthcare sector depending on your view of the economy. 

Even once you make a sector or industry choice, how you drill down into individual funds makes an enormous difference. Biotech ETFs, for instance, had returns in January ranging from 0.44% to -1.93%. The reason is that various biotech ETFs tackle different parts of the market: some focus on large established biopharmaceutical companies, and others focus on smaller, more nimble genomics plays.

The point is that investors have important choices to make with ETFs. They can buy full market exposure with a few funds and be done with it.  They can make broad size/style tilts and be satisfied. Or they can drill down to very specific industries and make specific allocations.  Even within each of those levels, they can make specific ETF choices that have a major impact on returns.

About The Ben Graham ETNs

As an aside, let me answer the question Jim posed in his last blog about why the Ben Graham ETN is structured as an ETN.  The first and primary reason is that it is an actively managed product.  There may be some quantitative metric behind it, but it is as active as it wants to be.  Since you can't run an actively managed ETF very easily right now, putting it in an ETN wrapper was the only solution for getting it launched.

The next logical question might be: why would an investor buy an actively managed equity ETN and take on the related counterparty risk, when they could just buy an actively managed mutual fund? It's a good question, and I'm not sure an investor should. 

But ETNs do have a major advantage for taxable investors, which is that it should not pay out any capital gains distributions.  Cap gains distributions cost actively managed funds 1-2% per year.  Is a 1-2% increase in after-tax returns each year worth the counterparty risk that accompanies an ETN?  Maybe, maybe not: but it's a legitimate calculus to consider.

SUMMARY:

I estimate Apple has sold roughly 6 million 3G iPhones with 3 weeks still left in the quarter. Q4 sales could hit 7 million, or more, with the aid of a couple factors.

1) July 11th, Apple introduced new model at 189 Apple stores and 2,000+ AT&T locations in the US, and internationally in more than 20 countries.

2) August 22nd, another twenty-plus countries launched.

3) September 9th, iPhone went on sale at Nearly 1,000 BestBuy stores.

Supply dried up at all points of sale after the first weekend. Apple stores received the bulk of new shipments as AT&T and foreign providers were strained for several weeks. About mid-August, the production ramp and demand levels showed signs of equalizing. It wasn’t until late August that AT&T stores began to have on-hand inventory since running out on the initial launch. Apple records iPhone sales when shipped to carriers opposed to Apple retail stores, which are recorded when sold to the end-user. As demand is starting to normalize along with Apple’s retail store inventory stabilizing, focus will shift to supplying the channel. This comprises of AT&T stores, foreign carriers from both the July and August launches, and BestBuy. With 6 million likely sold thus far, sales should surpass 7 million by the end of the quarter from replenishing the channel and consumer sales.

NET APPLICATIONS OS MARKET SHARE MODEL:

Net Applications estimates OS market share from internet usage data. Since the 2G iPhone supply dried up in May and June, the installed base essentially remained static at 6.12 million units during that time. As measured by Net Applications, iPhone OS share remained steady at around 16.5 bps for those two months. That equates to roughly 370 units per basis point of market share. Using the share data since the 3G launch, unit sales of the new iPhone model can be estimated.

There appears to be some aberrations in the Net Application survey data as the huge spike in mid-August would suggest. The spike causes the sales estimates to accelerate rapidly and then flatten. If the share data were more normalized, weekly sales estimates would be smoother. Thus, the weekly estimates are volatile, and likely not accurate. Yet, the data probably does give a decent estimation of cumulative sales since the launch.

Here are a several thoughts regarding the spike in web usage seen the week beginning August 8:

1) Certainly the weekly run rate has slowed, as with any new product launch, but definitely not to the degree that the OS share data depicts. Larger proportion of early sales (1st 30 days) versus late sales (2nd 30-days) whereby internet usage is highest right after purchase and fades. Thus, early iPhone purchases caused pronounced acceleration and when usage faded, late iPhones with heavy usage don’t offset the early decline since it’s a smaller portion.

2) There is cannibalization of 2.5G iPhones from upgrades. Original iPhones that become inactive from new 3G purchases won’t increase market share measurements. Of course, some legacy iPhones are sold or passed on.

3) Up until early August, supply was sporadic at Apple retail outlets, and virtually non-existant at AT&T and international carriers. A good number of orders were placed at AT&T outlets, but customers had to wait for their shipments to arrive. It’s possible than many customers who placed orders in July received them at the begging of August when supply firmed, leading to higher web usage.

AFB IMEI NUMBER TRACKING MODEL:

Members at the AFB board have been collecting IMEI numbers and recording those in a Google spreadsheet. The theory is that the IMEI numbers follow a consecutive sequence, and tracking them can reveal the number of units produced so far. The highest IMEI number submitted so far points to 5.604 million from a iPhone purchased on August 30. Fortune’s Elmer-DeWitt wrote about this approach September 1st. This model corroborates the Net Application data and my calculations, suggesting unit sales have approached 6 million.

STREET Q4 ESTIMATES:

These are the latest estimates that I have been able to find, however they may have been revised since.

Piper Jaffray- 4.5M

Credit Suisse- 4.2M

Pacific Crest- 3.5M

Financial Alchemist- 7-8M

CONCLUSION:

Apple’s 3G iPhone appears to be selling ahead of Street estimates which may provide an opportunity for upside surprise when Apple reports Q4 results in October. Even though demand has cooled from launch day, Apple has to supply the channel which suffered an inventory drought for over a month after the en model was released. Domestically, iPhone shipments will go being going to over 2000 AT&T outlets and nearly 1000 BestBuy locations. Abroad, some 50 countries will have their inventories replenished. This should give a boost to iPhone sales as the quarter comes to a close.

Looking at the OS market share data, the focus should be on cumulative units, as opposed to weekly change. I believe that some aberrations in the measurement may have lead to that abnormal spike seen at the beginning of August. This causes the subsequent weekly changes to show a pronounced slowdown which is likely exaggerated. We also must remember that some 3G models replace legacy iPhones which will not increase market share numbers, thus won’t account for new 3G unit sales.

The number to watch will be cash flow when Apple reports. The iPhone has the potential to significantly boost cash earnings. I discussed this implication back at the end of July- Apple’s Cash Earnings. On a cash earnings basis, Apple is very cheap at current levels.

Disclosure: none

EQUITY VALUE MULTIPLES:

Let’s take Apple (nasd:AAPL) for example: Price =  $172.55, Cash/Share = $23.45, FY09 EPS Estimate = $6.06. The forward P/E is 28.5x. The  incorrect computation is to subtract cash from the share price before dividing by expected EPS: $172.55 - $23.45= $149.10 / $6.06 = 24.6x. The rationale people give for making this mistake is that one share of Apple represents $23.45 of cash and a business that generates $6.06 in EPS, thus an investor can purchase the earnings stream for $149.10.

Here’s the issue- the cash balance contributes to earnings in the way of interest income. Without the cash stockpile, EPS would be lower. One must not assume that Apple’s FY09 EPS will be $6.06 without interest income, thus a higher multiple should not be assigned on the basis of its high cash/share. In FY07, Apple earned $647 million in interest from its cash holdings, which totaled $15.4 billion at year-end. In per-share terms, interest income contributed roughly 51 cents to Apple’s reported FY07 EPS of $3.93. Apple’s P/E multiple based on FY07 EPS is 43.9x. Without interest income, EPS falls from $3.93 to $3.42, and subtracting cash from share price, Apple’s historical P/E is 43.6x. That’s roughly the same as the multiple calculated with cash included in both price and EPS. The common mistake is not subtracting out interest income from EPS while taking cash out of the share price. Therefore, it’s incorrect to subtract cash from one figure without taking it out from the other figure as well.

Since P/E ratios represent income that includes interest income, the conversation of cash/share is inappropriate, as it has no bearing on value, nor multiples in that regard. It’s incorrect to assert that a firm’s P/E multiple is actually lower because it has a relatively high cash/share, and that one should consider cash/share in tandem with P/E ratio. The cash/share is accounted for in the P/E ratio because it’s a part of the “E” or earnings, which includes interest income. The cash balance is the present value of future interest income, thus the two are the same.

ENTERPRISE VALUE MULTIPLES:

In instances where EBIT or EBITDA figure (Earnings before Interest, Taxes, Depreciation, Amortization) is used in a price multiple, then cash holdings should be considered since interest income (expense) is not captured. Thus, a P/EBITDA multiple makes an incorrect comparison since cash & debt aren’t included in the value of the denominator but are in the share price, or market value of the equity. To properly compare EBITDA, one should use enterprise value, or EV, in place of share price, or P. EV is the market value of the equity plus value of debt minus cash. Therefore, the multiple becomes EV/EBITDA. Cash holdings are excluded from the value figure, numerator, as well as excluded from earnings stream, EBITDA, in the denominator. 

CONCLUSION:

To calculate multiples correctly, one shouldn’t include components in the numerator without also including in the denominator. If one is computing P/E multiple, then cash/debt needs to be ignored because those values are captured in the EPS. If one is computing EBITDA based multiples, then EV instead of P, is the correct input for the numerator. Since EBITDA doesn’t account for interest income/expense, then it would be much higher for a debt-laden firm. If share price, P, were used instead of EV, then the numerator would be too low resulting in too low of a multiple. Adding debt to arrive at EV, increases the numerator to coincide with the exclusion of interest expense increasing the denominator as well.