Western Digital (WDC):

Western Digital (WD) is the largest (by units shipped) maker of hard drives (HDD ) in the world. They make hard drives used in desktops, laptops, and net books as well as portable hard drives (last year desktops 38% of sales, all others 62%). They are known as one of the lowest cost producers in the business. Even though they shipped more units than their largest competitor, Seagate (STX), last quarter, WD had lower revenues as they have a lower average sell cost per unit. The industry as a whole is experiencing a cyclical high in margins and extremely strained capacity thanks to larger than expected demand in the past year which is expected to continue throughout 2010 resulting in very tight inventory (2 weeks for WD). The reason I believe WD is undervalued is due to the fact that most investors are expecting margins to fall back to normal levels. Yet as I show later, even if this happens due to changes in the industry, WD would still be considered a buy.

Industry considerations:

The HHD industry has become much more consolidated in recent years going from 8 major players 10 years ago to only 5 today. Based on last quarters shipment numbers the breakdown is as follows
[img]http://www.*********************/uploads/th.1c79ff600e.png[/img]
And for those who care, the HHFI is 2451, indicating a strongly concentrated industry. What this chart leads me to believe is that:
1. WD is gaining market share as this is the first quarter that they have been the leader in units shipped
2. Industry margins have undergone a shift and will no longer be fluctuating on the old cycle but on a new higher cycle.
In the past whenever the industry margins grew too ft, the many HDD manufacturers would ramp up CapEx too far, causing margins to become depressed over the next few years. Now however this will no longer happen as severely. The industry is more concentrated and the industry players are more open with respect to total CapEX. Here is an except from the CFO of WD during the latest conference call when asked about industry margins

“Okay, yes, we were encouraged by Hitachi's results and we think it's the level of transparency they are providing is a positive for the industry. And we also think that level of transparency will also be beneficial to the investment community as the behavior of the people now hold somewhere close to around 80% of the total market. And so everybody has got the information available to them in order to make their decisions relative to capacity additions and then once they have the capacity additions in place, they have the information available to them relative to capacity utilization.”

I believe a main reason WD is undervalued has to do with the expectations that WD margins will some come crashing down like they have in the past, and I believe this is false. Another threat to WD seen by many analysts is the shift to solid state drives.

HDD vs SSD

A lot of people and industry analysts seem to be under the impression that solid state drives (SSD) will soon kill Hard disk drives (HDD). They point out that SSD is so much faster than HDD, which is true. Yet SSD is also much much more expensive. For example a quick google search shows a random 32g SSD sells for $110 whereas a quick check shows you can get about 2 terabytes of HDD storage space for the same cost. I have read that this is only temporary and cost effective SSD is right around the corner yet that is also what I have been hearing since first following this industry is 2007 and nothing so far has led me to believe that we are any closer than before. In fact, since that time, WD and Seagate have both come out with hybrid HDD that are mostly solid state yet possess a certain amount of SSD for when you need to access data very quickly.

As well, companies that might eventually pose a threat to WD from the SSD world such as Sandisk have drastically cut down CapEx from 260 million in 2007 to 60 million in 2009. This gives less credence to the argument that they are getting ready to push into basic computers and basic external drives, WD’s bread and butter.
Finally, when SSD takes over in the future (it may be a long time off, but it will happen), WD will not be sitting still. Spending over 450 million in each of the past 3 years on R&D (464 million in 2008, 509 million in 2009, and 456 million through the first 9 months of 2010), as well as currently sitting on over 2.8 billion in cash, they have the resources and the expertise to quickly develop their own SSD technology. They just won’t bother until SSD becomes a cost efficient technology.

Management

The current CEO is John F. Coyne. He has been with the company since 1983 working his way up through different divisions within WD. Management has a good reputation for being conservative with their estimates and doing deals that make sense for shareholders. Looking over their last large deal, the 1 billion acquisition of Komag, we can see that they were able to complete the integration with lower costs than expected. In that conference call, management warned of lower margins for the next few quarters as a result of the deal. As you can see by looking at the excel file and the numbers for 2008, this never materialized. As well management warned of increased R&D cost of 20 million a quarter for the year, yet 2008 R&D spending came in at only 50 million higher than the previous year. Management also promised long term improvements from the deal of 250 bps to gross margin. Again from the excel file we can see that 2008, 2009 and the first 3 quarters of YE 2010 have experienced higher gross margins than 2007.

WD also has a good track record of not taking one time charges every quarter like their rival Seagate. I used to own shares in STX but became disconcerted with the constant restructuring and one time charges that appeared every single quarter. Eventually I gave up and sold my shares and bought WD instead because they didn’t seem to experience this problem. After following the company for several quarters and listening to the conference calls I also began to appreciate the conservativeness of management and their frankness (see above quote about industry CapEx). Only 1 year in the past decade has restructuring charges (2009, everyone restructured).

WD has strong corporate governance practices. The board is composed of 10 directors, 7 of whom are independent. The 3 remaining directors are 2 former CEOs of WD and the current CEO, John F. Coyne. The role of chairman and CEO is always separate and all the committees (audit, compensation, governance) are composed of independent directors. The entire board is up for election annually. All of these are consistent with best practices.

The one place for concern is insider sales. In the last 6 months insiders have sold half of the shares they held, although this is only .0025 of all shares held. This could be due to the fact that management is often rewarded with stock grants rather than options. As well the shares have recently traded at their all time high of 45. It would be nice to see some buying soon though around the 31, 32 level.


Valuation:


To make the valuation model I assumed that the past 5 years (9 months for the past year) represent normalized earnings. Based on my comments above about the change in the competitive landscape of the industry I believe this is more accurate than a 10 year model. Examining the past 5 years we see that the gross margin has been about 20% with a high of 25% this year and a low of 16.46% in 2007. So I used this as a base to get gross profit.

Sales were estimated extremely conservatively. They have been growing at a CAGR of 20% over the past 9 years. Instead of using this I used total industry HDD shipment expectations of 5.4% yearly growth over the next 4 years. I don’t believe that WD will experience sales growth this low as Asia makes up a larger and larger share of their customers (asia contributed 33% of total sales in YE 2007 vs 50% in YE 2009) and they should continue to experience rapid capital growth, I figure I will be conservative.

I kept R&D at a constant % of sales.

I estimated SG&A would be 235 million this year. I regressed the relation between sales and SG&A which resulted in a slope coefficient of .011. I also calculated SG&A as a % of sales in each year and it has been slowly declining but it makes more sense to use the slope rather than the trend of sg&a as a % of sales in my opinion since you can only cut so low before economies of scale taper out. With this I took forecasted SG&A for year 2010 and added [marginal sales each year x the slope] to come up with incremental sg&a spending for the forecast.

Interest expense is expected to be minimal since the giant surplus of cash and taxes are estimated to continue at the current rate of 9%

Terminal growth is expected to be 4% a year, in line with world growth. If I were modeling a mostly North American company I would use lower long term growth forecasts, but WD is increasingly a global player and now Asia accounts for half of revenues. Since this region is expected to grow much more quickly for the foreseeable future than Europe and North America, I think 4% long term growth is a fair assumption. I expect to reach the terminal level of growth after 3 years. This may be quick ands the company may still be growing rapidly 3 years from now, but id rather be caution and not estimate too far out. The discount rate is a standard 10%, the minimum return I would expect from my investments. As a note, I am not crazy about using betas and WACC all that much as stock price vol doesnt necessarily equate real risk. In any case the beta is 1.5 and the capital structure is almost all cash so the cost of equity using a 5% market risk premium and the 10 year as the rfr is about 10.8% which is lower than the 12% higher end I will use in my calculations (next paragraph).

Using these numbers we get a value of 16.4 billion. Changing the terminal growth rate to 2% we get 13.1 billion. Changing the discount rate to 12% and the growth rate to 2% we still get a value of 12 billion. Changing all those numbers and putting the margin at the cyclical low of 15% going forward we still get a value of 7.16 billion, substantially higher than the current market value – cash.
I also constructed a FCFE model. Under this I assumed ROE would be 15% going forward for 3 years and then drop to my long term growth rate of 4%. ROE for the last 5 years has averaged 43% so I do not think 15% going forward is a crazy estimate. I think 15% is low enough under the average that I have accounted for the low ROE next year due to the giant cash position.

Given that assumption and the fact that they retain almost 100% of their income, the growth rate should match ROE. Using this year’s FCFE (9 months actual and 3 months forecast by management) of 1188 and my assumptions, along with a 10% discount rate, FCFE comes in at 18.1 billion. Using the average FCFE rate of 715 million over the past 5 years and the growth assumptions we get a valuation of 10.84 billion.
Considering the market cap less cash is only 4.84 billion WD seems like a screaming buy. Under the most stringent assumptions it is still selling for much less than what it appears to be worth.

Uses of cash:

Since the cash on the balance sheet is such a large part of market cap it is worth talking about what management intends to use it for. One option would be to begin issuing dividends but because of their tax rate it would be disadvantageous to do that. The dividend would be taxed at a much higher rate than the 9% WD pays. Another option is to buy back shares. Management has shown a small willingness to do this in the past and has 466 million remaining on a repurchase authorization. I expect they will use this up shortly. The final option is to acquire competitors and new technologies. Last spring they spent 65 million to acquire SiliconSystems, a maker of SSD. It is due to acquisitions like this that I am not worried about WD falling behind the technology curve. 2 years ago they acquired a large player in the same HDD industry, Komag for 1 billion. This acquisition went smoothly with all synergies realized. I would find it hard for WD to duplicate this move however, as the industry is already quite concentrated and any further moves might bring on the scrutiny of regulators. Therefore I predict that the main use of cash will be for CapEx when needed (about 600-700 million a year) with the rest being returned to shareholders via share repurchases.

Final thoughts:

It seems to me that WD is a clearly undervalued company. The market cap of 7.66 billion includes more than 2.8 billion of cash and the company is still gaining market share in a growing industry. Even though the HDD market can be considered slightly commoditized at this point, the recent consolidation and clarity with regards to CapEx will lead to continual high margins over the next few years. I believe the true value is close to 15 billion, or close to double what it is currently trading at.

I have the excel file I used up on my website at www.freenpv.com

if anyone could tell me how its possible to load word files complete with charts onto a website or a post like this I would appreciate it. Loading it up to freeimagehosting every time is going to get pretty annoying in the future

I recently took a look at WD and STX and agree WD is very cheep right now. I think you embedded some important basic points in your analysis which stuck out to me and are part of the fundamental attractiveness of this idea.

1. Larger scale that all its competitors. This should help WD stay the low cost producer, which is the most important factor in producing a commodity product. WD has grown to this point from the 5th largest player not so long ago which indicates excellent operations management.
2. Favorable industry dynamics. With only five real competitors in the hard drive industry, the makers have above average leverage with their suppliers and below average pressure from customers due to buyer fragmentation (it still is a commodity product, so this may be less important)

im also looking for critiques and criticisms of my work. Id like people to challenge my assumptions if at all possible, unless you all agree with most things I wrote.

1. why is SG&A constant? that doesn't make sense to me on the face. what does the past n years' balance sheets say about this assumption?
2. discount rate seems to be pulled out of thin air. tech companies i'd expect to have higher cost of equity. what assumptions did you use to calculate the WACC (did you calc this/estimate ranges for it?)?
3. why 4% (2%) for terminal growth rate? population seems like a pretty poor estimate for long term growth.
4. why terminal growth rate in only 3 years? that seems VERY soon to be expecting stable long term growth, right?

those are just a few things that jump out atm.

1. SG&A is constant because there is a pretty weak relation between sales and SG&A. They have tripled sales in the past 5 years yet SG&A costs have only gone up 52% in the same time frame. Looking at the excel the relation is so weak that modeling it off sales doesnt seem to make a ton of sense. Thats not a great excuse for not coming up with something better. But even if I assume it grows at 1/3 the rate of marginal sales with the low sales growth I used for the next 3 years it will only add 2% a year to SG&A. So maybe I should change it but I doubt it will make a large difference.


2. Discount rate is semi pulled out of thin air. Im not crazy about using betas and WACC all that much as stock price vol doesnt necessarily equate real risk. In any case the beta is 1.5 and the capital structure is almost all cash so the cost of equity using a 5% market risk premium and the 10 year as the rfr is about 10.8% which is lower than the 12% higher end I used anyways. 5% market risk premium is a higher mrp than iv found in the past (think last time I ran excess returns over the past 60 years it came out to something like 4%) but seems to be what people use in general.

3 +4. I used these numbers to be conservative. I do not think that they represent the growth prospects for this company at all. I doubt that sales will grow slower than pop. levels so that is a base level I felt comfortable using. but also the industry numbers I saw only predicted total shipment growth in the 5-6% range for the next 5 years. So I just chose the lower number to not overestimate.

Also 3 years is soon to enter terminal phase but I didnt see any huge changes coming to this company in the next few years so I figured id estimate a few years out and then just assume constant slow growth about the same rate as the economy. What would you have done?

Thanks for the critiques

good post.

what i would have done:

1. what you say here. graph SG&A vs. a few things that you are predicting and determine the 'best' way to anchor it to something. sure it wont make a large difference but it will add to the credibility of your write up.

2. even after the recent crisis you get just 4%? i haven't done it in like 4 yrs but it was 7% when i was in bschool. shocking to see it drop nearly by half in 4 yrs lol. anyways, i'd try to be as accurate as possible and then draw a wide range around the assumptions. what i'd really do (if i were doing this for a high price client or something similar), since this is such a crucial assumption, is actually path model this in matlab, cb, or even excel (though its not as flexible). that said, the marginal improvement would surely be minimal (though it would, again, add a great deal to the credibility of the report).

3. i probably would have tried to be as "accurate" as possible, again noting where i made assumptions, and then, from there, back down to far more conservative #s. i'd probably end up with 3ish% as the terminal growth rate as that is what i've typically assumed the economy to grow at long term. but for valuation purposes, the conservative bent surely paints the picture.

fwiw, we're actually now considering WD and doing our own due diligence on it, the management team etc.

i guess now would be a good time to point out that i think there should be more w.r.t. the management team in the writeup. who are they? do you like them? if so, why? what initiatives do they have going on? what if they fail? and lots more, etc. we've found companies that looked to be SCREAMING buys that we passed on b/c of the management team so i'd def focus more on that.

i'll post more if i think of any...

all,

ahnuld has laid out a really solid baseline framework for company assessment/valuation. i see many on this board constantly thinking/posting about stocks and i disregard like 95%+ of those since they don't typically involve any discernable level of thought/analysis.

ahnuld, one more thing i'd add is a balance sheet assessment. i.e. we like to know what our "safety net" is in terms of valuation.

take the BS, cut receivables by some % (like AR*.75), cut goodwill by like 50% and 90%, cut intangibles by some % and cut net PPE by some %. add the rest up at full value, then subtract all debt/payables etc. at full value. divide the result by the # of shares and report your assumptions and the resulting value. that is a decent proxy for your valuation "safety net" if you wanted to gekko the company ;-)

ahnuld, thanks for this thread. I haven't read it in detail yet, but really, these are the kinds of threads that people should be thankful for in this forum.

Great writeup Ahnuld. Thanks for this.

This is the type of thread this forum was made for.

The only suggestion I would have is to list the various rates you have used, the values you chose, along with the rationale, in one convenient spot towards the top of the relevant section.

WACC = [blah blah blah]
Discount rate = [blah blah blah]
etc.

ok, to fix problem 1. I ended up regressing the relation between sales (ind) and SG&A (dep) to get a slope of .011. I also calculated sg&a as a % of sales in each year and its been slowly declining but it makes more sense to use the slope rather than the trend of sg&a as a % of sales imo.

Anyways, so I estimated the sg&a for this year and then added the marginal sales year year x the slope to come up with incremental sg&a spending for the forecast. I think that works.

point 2. I maybe be confused but if anything the crisis should have lowered the market risk premium, no? I mean mrp = avg stock returns - avg bond returns right? so in that case since stocks have preformed so poorly in the past 3 years and bonds so well the rate would probably closer to 3.5% or something by now. anyways, last time I did it was for a ptest for an internship and because the equity risk premium was so low I kept getting a ******edly high value for my model and the guy basically told me they just use about 5%. Def. not scientific and if I had to do this for a client I would probably focus a bit more on it but this is just for me and I think 5% is a comfortably conservative number. But I should write a paragraph explaining this and why I chose 10% (which is probably pretty close to the wacc given mrp is below 5%) and 12%.

3. I like 3% too when valuing a domestic company. But WDs breakdown of sales is slanting more towards the developing countries who should experience higher growth than us for awhile. so 4% is ok imo, but I could see an argument for 3%. Maybe if I didnt project terminal value to be so soon, say in 5-10 years 2-3% terminal growth would be better.


thats a good point on management. Intuitively I know why I like them but I forgot to write anything about it. Def. a good idea.


anyways, thanks for the thoughts, would appreciate any more you have. ill be gone for the weekend but back sunday night

question: isnt this more relevant when investing in a financially distressed company? I mean if we are looking at a company with zero debts and extremely positive cash flow im not 100% sure of the value of this.

another company ill do a writeup of soon, friendman, is definitely trading as if it is distressed and more of a balance sheet play, so I understand the merits of that sort of analysis. Just not sure if it is very relevant to the investment decision making process in WD's case.

The bulk of "desktop" hard drives - where do their sales come from? Business or retail (home)?

request writeup of ceu?

Please do.

This forum needs more threads like this. I might comment after I've read more about it.

I use both DCF and EPV valuation methods to come up with a range of estimates for a given company. With some quick calculations, I came up with a value of $38-$40 per share.

I used a discount rate of 15%, Owner Earnings of 430m, and a growth rate of 15%. For EPV, I used a normalized income of 833.5m.

Going back a year, there has been a decent amount of insider selling at prices ranging from $30 to $40. I'd have to dig into the SEC filings to further refine my estimates.

I agree with many here - awesome post. I started my own value investing blog where I post similar stock writeups & analysis (not sure of the rules for posting a link to it?)

If we start making this a trend, I'll try to post some of my ideas here as well.

very interesting.

macro analysis about the firm/market sound good. at least for me, it was tough to follow the valuation portion because some of the numbers were guesstimated and you obviously arrived at final figures on your own. depending on who the target audience is an appendix with calculations and your assumptions behind growth rates could be useful.

would you say its good practice to add BS assessment to every company or is it insignificant for financiall sound firms (these guys have a ton of cash - do we need this? also seems like it would be a fairly biased metric if we are arbitrarily cutting all these BS items).

ty for this write up, i am reading some textbooks and hope to learn enough so that i can one day evaluate a company like you did here.

Yeah I'm interested in this kind of assessment as well.

One thought: Yeah it makes sense to choose the parameters as accurately as possible, but it seems like that might be pretty hard. Even if you have the best guess out there, it still might not be particularly convincing since the spread of possible values might be very large. It makes sense to me to present what you consider to be your "best" numbers, but I would definitely run an absolute ****load of other regressions, etc. for sensitivity. It's obviously really important to know whether your results hold under a wide range of assumptions, or only a narrow range (at which point the accuracy of the parameters becomes much more necessary).

I could find the exact numbers if you wanted but STX focuses more on the high margin business users and WDC the retail side (laptops, desktops and portable harddrives)

management writeup is now included

Great Writeup. What was CEO's path to top? Is he more of a tech guy or a sales guy or whatever?

So are you loading up at 30?


My concern would be the disruptive effect of cloud computing on the HDD market.

That argument came up with a friend of mine. we decided that since wdc focuses on the lower end of the amrket (retial, not business) they are less at risk of cloud computing than stx.

and I got long at 20 in late march 09. Probably going to pick up some more here, but im debating between that and waiting for earnings in 3 weeks since its already a decent sized position.