IBM - International Business Machines

[kevin4u2]

Your absolutely correct but it doesn't change those results.  It is also why I discussed the return on Total Capital and did not mention that ROE over the last ten years has risen from 25.7% to 132.7% last year.  That is the effect of leverage.  The increasing margins I presented do not depend on the capital structure. 

 

Everyone is focused on the lack of revenue growth at IBM but they miss the mark.   

 

Management has done an excellent job with the company.  Here is how:

 

Ten year revenue growth per share 6%.

Ten year earnings growth per share 13%

Ten year dividend growth per share 19.5%

Over the last ten years, Operating Margin is up 10% to 26%

Over the last ten years, Net Profit Margin is up 7.7% to 17%

Over the last ten years, Return on Total Capital is up 16.3% to 34.5%. 

 

Operationally, that is incredible performance and everyone is all bent out of shape over a lack of revenue growth.  They generate much higher returns from their assets than ever before.  I would love to see the average armchair analyst demonstrate that type of performance.  Perhaps IBM knows how to invest capital, and they know how to do it very well. 

 

How many businesses earn that level of return on capital?  How about Amazon, what is their Operating Margin? 5.5%. Net Profit Margin? 0%. Return on capital? 0%.  Why does Amazon get a pass? 

 

As for the CAPEX argument, let's look at some facts.  Both companies spend roughly the same amount on CAPEX as each other, at around $4.5 billion.  Everyone says IBM doesn't invest in their business.  We'll no, they just happen to generate over $20+ billion per year in CF, they invest the same amount as AMZN and still happen to have $15+ billion in free cash to pay dividends and buyback shares.  AMZN generates $4.5 billion in CF and invests the same amount in CAPEX, makes no money doing it, yet they are awesome. 

 

Amazon is great for social good, we all enjoy the benefits.  The problem is their business model benefits consumers and not investors.  It's basically a free service.  To those who are shareholders, I send my deepest thanks.

 

Good perspective. However let me remind you that all of the above was achieved with their debt to equity going from 0.68 to 3.44. With LT debt going from circa 15billion to over 35billion.

Now i beleive in the investment thesis and have a long position etc. and I agree that for a company with such stable and recurring cash flows in the midst of a transition, and with such historically low interest rates it was arguably the right thing to do for shareholders. That said the company has levered up significantly and once revenues start growing again, will need to shift gears to deleveraging the balance sheet rather than continue share buybacks. I suspect thats what they are going to do as well.

[vinod1]

Longtime Lurker, first time poster.  Wondering if anyone familiar with IBM would be willing to help me out?  I had to write a report for an interview and I'm hoping someone might be able to review for me.

 

I think if you post it here you would get feedback. I would be happy to help.

 

Vinod

[vinod1]

1. Is it an above average business?

2. Is it selling at a below average price?

 

If you put it like that, I disagree with 1.

 

Edit: to elaborate, I believe that IBM has at most 60% of sales in above average business that is not growing. The rest is potentially a melting ice cube.

 

Edit2: to be complete, I will also disagree with 2. :) If I put GF debt and pension liabilities into EV, IBM is trading at below 5% E/EV.

 

Jurgis,

 

I hope you would reconsider participating in this thread. You have a different slant at most things so that would be a very helpful perspective.

 

1. Software is 25% and services are nearly 60% and nearly 70% of services revenue tied to very long term contracts. So taking just the long term contracts and software that is 2/3 of IBM revenue that have some sort of moat. I would speculate that a higher proportion of the earnings say 75% to 80% comes from these moaty parts.

 

2. I do not think it makes any sense to look at financial services debt when calculating enterprise value. Debt is to banks as steel is to a car company, an essential ingredient in the services they offer. So I would not look at enterprise value for a financial company.

 

Unfunded Pension liabilities are conservatively about $15 billion or around 1x earnings.

 

So you can argue that it is roughly 11x earnings or a 9% earnings yield.

 

Vinod

[Jurgis]

Vinod1,

 

Just a quick answer.

 

1. Hmm, I actually think that mainframe HW has bigger moat than SW. Clearly Rational is bleeding - which I kinda would have expected. I think services are hard to quantify - I agree some of them are sticky, others less so. Since services is a large part of revenues, we might have tough time coming to a common view. :)

 

2. I am not sure what numbers you are using:

2014: 168 market cap + 15 pension / 12 earnings = 15 P/E

P/FCF is similar though a bit higher

2015q1 annualized: 168 market cap + 15 pension / 2.4x4 earnings = 19 P/E

FCF - they reported Free Cash Flow (Excluding GF Receivables) 1B in Q1.

 

I will also post what I sent to cubsfan in private:

 

I think the discussion was somewhat exhausted anyway: I see bull points, I disagree with them partially, I don't think I'm gonna change anyone's mind and they are not changing mine. :) I agree that it is possible that IBM will be a good business and good investment, but I think it is not good enough and not cheap enough for me. It's quite possible that bulls will be right, I'll be wrong. I have no issue with that especially since I hold BRK and FFH.

 

Good luck, was fun and all that. :)

 

Hopefully this clears up and somewhat wraps up things. I really don't want to "fight" on bear side and get attached to that being "my" side, so I'd rather let other people discuss things for a while at least.

 

Thanks for understanding.

[vinod1]

Vinod1,

 

Just a quick answer.

 

1. Hmm, I actually think that mainframe HW has bigger moat than SW. Clearly Rational is bleeding - which I kinda would have expected. I think services are hard to quantify - I agree some of them are sticky, others less so. Since services is a large part of revenues, we might have tough time coming to a common view. :)

 

2. I am not sure what numbers you are using:

2014: 168 market cap + 15 pension / 12 earnings = 15 P/E

P/FCF is similar though a bit higher

2015q1 annualized: 168 market cap + 15 pension / 2.4x4 earnings = 19 P/E

FCF - they reported Free Cash Flow (Excluding GF Receivables) 1B in Q1.

 

I will also post what I sent to cubsfan in private:

 

I think the discussion was somewhat exhausted anyway: I see bull points, I disagree with them partially, I don't think I'm gonna change anyone's mind and they are not changing mine. :) I agree that it is possible that IBM will be a good business and good investment, but I think it is not good enough and not cheap enough for me. It's quite possible that bulls will be right, I'll be wrong. I have no issue with that especially since I hold BRK and FFH.

 

Good luck, was fun and all that. :)

 

Hopefully this clears up and somewhat wraps up things. I really don't want to "fight" on bear side and get attached to that being "my" side, so I'd rather let other people discuss things for a while at least.

 

Thanks for understanding.

 

Jurgis,

 

In 2014 they had a Loss from discontinued Operations net of tax of about $3.7 billion. So I think using 2014 numbers without this adjustment would underestimate the earnings power.

 

My valuation is below:

 

http://vinodp.com/documents/investing/IBMValuation.pdf

 

IBM is a small position for me and I purchased it below $155 and would add to it if it goes below that level. It is not a punch card type investment for me, I just think the risk reward is compelling at around $150.

 

Vinod

[cubsfan]

Vinod - very nicely done. Thank you for posting.

[rishig]

My friend Jana (https://janav.wordpress.com/2015/04/24/buffett-to-blogging/) recently presented his case on IBM to our little club of value investors in the Bay Area. Check out his presentation: https://janav.files.wordpress.com/2015/04/ibm.pdf

[Straddle]

Vinod1 - Thanks for your summary. Very good written.

[orthopa]

I wish I had something substantial to add. Just want to say thanks to the great minds in this thread. I learn more everyday. Long IBM btw. For all the reasons presented above.  ;D

[fareastwarriors]

 

Can Mr [bill] Gates be credited with getting Mr Buffett over his famous reluctance to invest in technology shares with the IBM investment?

 

“No, IBM became less of a technology company,” Mr Gates says. “It’s really sad. It turns out, at least so far, Warren was wrong. Even with the enterprise customers, the cloud has not been saleable for them. I’m biased. IBM is a wonderful company but because I competed with them for many, many decades I have to say, I don’t see their future as brightly as people who are long on the stock.”

 

 

 

http://www.ft.com/intl/cms/s/2/e7cae5e0-ef24-11e4-87dc-00144feab7de.html#slide18

[TwoCitiesCapital]

http://finance.yahoo.com/video/why-ibm-not-apple-warren-103900597.html

 

Buffett's thoughts on IBM

[LowIQinvestor]

You guys saw that Warren bought more IBM in Q1?

 

Completely agree with him that Watson will be an incredible platform

 

IBM is a buy today. I don't care what Yellen says :)

[LowIQinvestor]

Hope IBM and Warren are picking up some shares today.

 

On sale at 10x earnings.

 

[valuecfa]

Curious if other investors think Watson will be a game changer for IBM over the long term (not a short term thesis)? IBM predicts over $1 Billion in revenue form Watson by 2018 (2014 article). Small compared to overall revenue, but long term it is an interesting revenue stream. Given the likely very high margin provided from Watson and other AI computers, it's interesting to read about the partnerships/apps that are collaborating with Watson. Also, any thoughts on the model they are using to market and generate revenue with Watson vs other companies (revenue share/subscription)?

[peter1234]

Watson =PR. With cloud computing, others can build better services.

[Palantir]

Curious if other investors think Watson will be a game changer for IBM over the long term (not a short term thesis)? IBM predicts over $1 Billion in revenue form Watson by 2018 (2014 article). Small compared to overall revenue, but long term it is an interesting revenue stream. Given the likely very high margin provided from Watson and other AI computers, it's interesting to read about the partnerships/apps that are collaborating with Watson. Also, any thoughts on the model they are using to market and generate revenue with Watson vs other companies (revenue share/subscription)?

 

No! Watson is a solution looking for a problem.....

[TheAiGuy]

Curious if other investors think Watson will be a game changer for IBM over the long term (not a short term thesis)? IBM predicts over $1 Billion in revenue form Watson by 2018 (2014 article). Small compared to overall revenue, but long term it is an interesting revenue stream. Given the likely very high margin provided from Watson and other AI computers, it's interesting to read about the partnerships/apps that are collaborating with Watson. Also, any thoughts on the model they are using to market and generate revenue with Watson vs other companies (revenue share/subscription)?

 

No! Watson is a solution looking for a problem.....

 

I agree that Watson is a "solution looking for a for a problem" but I don't agree that it's strictly PR or particularly "game changing". It looks like a branded consulting solution to facilitate machine learning in non-technical settings, possibly in an effort to sell other products.

 

We use a lot of machine learning tools in a research setting (I'm a researcher professionally) and they're very powerful but require 1) technical expertise and 2) content specific knowledge. Finding people with content specific knowledge and technical expertise is incredibly difficult outside of a few fields (computer science, academia, etc.). Presumably, it is difficult to find a good tax attorney who is also a skilled programmer and proficient in applied math. However, if you did have a good tax attorney/ data scientist, there are presumably a good number of questions you could ask this person about the tax implications of a particular merger.

 

Traditionally, these problems have been solved with large amounts of human labor and without the math. It makes sense for large expensive transactions, but doesn't make sense economically for smaller decisions. For example, it is expensive to organize a large team of experts to pour over histories of patient outcomes and recent medical trials to come up with a course of treatment for a single patient. You could make this process cheaper by having me do it with fancy math, but my salary would be expensive and the process would be too slow.

 

Enter Watson. If it is an AI/speech recognition front end for machine learning built on a custom data set, then that seems technically feasible and potentially quite valuable. You would have to custom build it on top of the customer's data set for each customers, but there is not reason why that shouldn't work or why JP Morgan or Partner's in health wouldn't find that valuable. Because it has to be custom built for each customer, it is unattractive for tech-companies without a history of consulting. So, basically, it is a taylor made product for IBM. I have no idea how large the market is; you need fairly large organizations that have realistically speaking, hundreds or thousands of unique (not repeated) questions a day. The organization needs to be large enough for scale, but the questions need to be small or else it makes more sense to employ a team of data scientists.

 

Competitors will never take on "Watson", but there will be (and already are) smaller solutions focused on a narrow subset of the addressable problems. For example, private wealth management is a potential avenue for this type of product, but there are several "robo-advisors" (e.g. Betterment) that directly compete.

 

 

 

 

[Jurgis]

For example, private wealth management is a potential avenue for this type of product, but there are several "robo-advisors" (e.g. Betterment) that directly compete.

 

OT. I am not sure private wealth management is a good example. IMO, there are two levels for wealth management:

 

1. Asset allocation. This is something that robo-advisors do OKish. There is no evidence that ML/AI can do it better. At least not current level ML/AI. In short, I know no data that would show that changing REIT allocation from X% to X+delta% would improve your results while lowering risk. Although I am sure Ray Dalio has a ton of people working on this.

 

2. Wealth management as hand holding. This is not doable until we get full AI. This is handling your client's kid education expenses, divorce, estate, not allowing them to sell during market crash etc.

 

I am not sure there is anything ML/AI can do that's in between 1 and 2. But I'd love to talk if you have thoughts. Possibly not on this thread. :)

[LowIQinvestor]

The great thing about IBM is that at today's price of around $162 you don't need any ground-breaking innovation to achieve a satisfactory long term return.

 

I would put 20% of my net worth in IBM today.

[TheAiGuy]

For example, private wealth management is a potential avenue for this type of product, but there are several "robo-advisors" (e.g. Betterment) that directly compete.

 

OT. I am not sure private wealth management is a good example. IMO, there are two levels for wealth management:

 

1. Asset allocation. This is something that robo-advisors do OKish. There is no evidence that ML/AI can do it better. At least not current level ML/AI. In short, I know no data that would show that changing REIT allocation from X% to X+delta% would improve your results while lowering risk. Although I am sure Ray Dalio has a ton of people working on this.

 

2. Wealth management as hand holding. This is not doable until we get full AI. This is handling your client's kid education expenses, divorce, estate, not allowing them to sell during market crash etc.

 

I am not sure there is anything ML/AI can do that's in between 1 and 2. But I'd love to talk if you have thoughts. Possibly not on this thread. :)

 

Agreed. Robo-advising isn't likely a good use of ML/AI (though, you run into people who want to use it for trading all the time, ala Renaissance Technologies). I meant more in terms of tax planning for high net worth individuals, where you might be able to do something like use natural language processing to better understand tax and previous court decisions for tax planning purposes. I'm making this up and I have no idea how feasible/expensive/ridiculous this is, but this type of thing certainly possible.

[Guest Schwab711]

Curious if other investors think Watson will be a game changer for IBM over the long term (not a short term thesis)? IBM predicts over $1 Billion in revenue form Watson by 2018 (2014 article). Small compared to overall revenue, but long term it is an interesting revenue stream. Given the likely very high margin provided from Watson and other AI computers, it's interesting to read about the partnerships/apps that are collaborating with Watson. Also, any thoughts on the model they are using to market and generate revenue with Watson vs other companies (revenue share/subscription)?

 

No! Watson is a solution looking for a problem.....

 

I agree that Watson is a "solution looking for a for a problem" but I don't agree that it's strictly PR or particularly "game changing". It looks like a branded consulting solution to facilitate machine learning in non-technical settings, possibly in an effort to sell other products.

 

We use a lot of machine learning tools in a research setting (I'm a researcher professionally) and they're very powerful but require 1) technical expertise and 2) content specific knowledge. Finding people with content specific knowledge and technical expertise is incredibly difficult outside of a few fields (computer science, academia, etc.). Presumably, it is difficult to find a good tax attorney who is also a skilled programmer and proficient in applied math. However, if you did have a good tax attorney/ data scientist, there are presumably a good number of questions you could ask this person about the tax implications of a particular merger.

 

Traditionally, these problems have been solved with large amounts of human labor and without the math. It makes sense for large expensive transactions, but doesn't make sense economically for smaller decisions. For example, it is expensive to organize a large team of experts to pour over histories of patient outcomes and recent medical trials to come up with a course of treatment for a single patient. You could make this process cheaper by having me do it with fancy math, but my salary would be expensive and the process would be too slow.

 

Enter Watson. If it is an AI/speech recognition front end for machine learning built on a custom data set, then that seems technically feasible and potentially quite valuable. You would have to custom build it on top of the customer's data set for each customers, but there is not reason why that shouldn't work or why JP Morgan or Partner's in health wouldn't find that valuable. Because it has to be custom built for each customer, it is unattractive for tech-companies without a history of consulting. So, basically, it is a taylor made product for IBM. I have no idea how large the market is; you need fairly large organizations that have realistically speaking, hundreds or thousands of unique (not repeated) questions a day. The organization needs to be large enough for scale, but the questions need to be small or else it makes more sense to employ a team of data scientists.

 

Competitors will never take on "Watson", but there will be (and already are) smaller solutions focused on a narrow subset of the addressable problems. For example, private wealth management is a potential avenue for this type of product, but there are several "robo-advisors" (e.g. Betterment) that directly compete.

 

I thought the point of Watson was highly technical applications. Watson is a learning wikipedia that is not connected to the internet. The anti-cloud if you will. I think there will eventually be numerous uses due to the enhanced security Watson provides.

 

Your provide an interesting perspective. IBM mentioned healthcare applications once upon a time and that always seemed like the perfect industry to me.

[Jurgis]

I meant more in terms of tax planning for high net worth individuals, where you might be able to do something like use natural language processing to better understand tax and previous court decisions for tax planning purposes. I'm making this up and I have no idea how feasible/expensive/ridiculous this is, but this type of thing certainly possible.

 

OK. This might work though it's definitely not straight-ML. You'd need what you said in above post: Watson'y tech + a bunch of experts + a bunch of work to get the stuff in and usable. Even then, it's not clear on who the customers would be, how much they would pay, whether ROI is worthwhile. And, yeah, I am also just guestimating.

 

The legal profession seems to be disruptable by AI/ML/Watsony things. Though probably at data/argument preparation only. I am sure there's enough human factor at the trials/hearings.

[portfolio14]

The great thing about IBM is that at today's price of around $162 you don't need any ground-breaking innovation to achieve a satisfactory long term return.

 

I would put 20% of my net worth in IBM today.

 

This is only true if the shift to cloud (i.e. IT becoming an utility) doesn't fundamentally shift the power in the value chain. e.g. The arrival of digital music shifted the power from the labels to the iTune and the consumers.

 

One thing that has puzzled me all along is, if the premise of cloud to corporates is lower TCO, someone in the value chain must be making less money. Technological upgrade is frequent, not unlike aircraft upgrade. So, it looks to me it's the cloud providers facing increasing cost but lower revenues. Don't the cloud providers look like airlines?

[Jurgis]

Don't the cloud providers look like airlines?

 

They look like airlines, but they hope to be more sticky. Some of them might be.

[TheAiGuy]

Eh,

The great thing about IBM is that at today's price of around $162 you don't need any ground-breaking innovation to achieve a satisfactory long term return.

 

I would put 20% of my net worth in IBM today.

 

This is only true if the shift to cloud (i.e. IT becoming an utility) doesn't fundamentally shift the power in the value chain. e.g. The arrival of digital music shifted the power from the labels to the iTune and the consumers.

 

One thing that has puzzled me all along is, if the premise of cloud to corporates is lower TCO, someone in the value chain must be making less money. Technological upgrade is frequent, not unlike aircraft upgrade. So, it looks to me it's the cloud providers facing increasing cost but lower revenues. Don't the cloud providers look like airlines?

 

I don't know that that is true -- that a shift to the cloud make IT a utility. A shift to the cloud obviates the needs to maintain and update your own servers, which is cap-ex intensive and difficult in its own right. But servers aren't the same as IT, you still need software (probably customized to your business) interaction with machines all over your business network. Maintaining servers is just one part of the value chain.

 

Also, how is the cloud like airlines? Airplanes have lifespans over decades, server farms have to earn their cost much more quickly. Also, I don't think server usage is as price elastic as airplane tickets, or as subject to discretionary cutbacks during a business slowdown. If you're Target, you can't tell your employees to process less data when business is slow but you can tell people not to fly (i.e something like email and teleconferencing acts as an imperfect substitute). You can restructure your business to lower the expense, but it becomes like restructuring your business to use less electricity.

 

Don't the cloud providers look like airlines?

 

They look like airlines, but they hope to be more sticky. Some of them might be.

 

It should be a sticky business. Changing core infrastructure in a company can be done, but its a pain in the ass and won't be done lightly. Note, I'm assuming that computing is a core business function to basically everyone.