Description

Computer Associates is a diversified software company specializing in systems and network management (i.e. monitoring large computer systems), storage software and security software. The company was founded in the 1976 by Charles Wang and dominated the mainframe management market. Throughout the 1980’s and 1990’s, CA was perceived as a very aggressive company, on many fronts, including acquisitions, customer relationships and accounting (see customer perception section). During the late 1990’s a new crop of best of breed software vendors as well as a rejuvenated IBM, began to take share from CA, as these companies exploited customer negative sentiment towards CA. In October 2000, CA underwent a dramatic transformation and shifted its revenue recognition model (and ultimately business model) from an upfront recognition ‘perpetual’ model to a ratable deferred recognition ‘subscription’ model. With this change, CA adopted a much more customer-centric approach (this was a watershed shift for CA as well as for the software industry) and allowed customers to structure contracts much more flexibly. This shift also dramatically reduced the incentive for salespeople to book ‘11th hour’ contracts and improved CA’s linearity, visibility and accounting transparency. CA also began a large management and board overhaul, which is nearly complete today (the CEO is the only person that is left from the earlier days and he only became CEO in 2000). In February 2002, both the SEC and the US Attorney’s office, began investigating CA for its accounting practices. Additionally, as part of the board overhaul, CA’s new board, led by the SEC’s former chief accountant, agreed to conduct an internal audit. In October 2003, the board announced that it had found questionable revenue recognition practices in March 2000 (see below for more details). With that finding, CA’s CFO (from the old regime) and three members of the finance department left the company. The board has said that it plans to conclude its investigation in the next 1-2 months. The SEC/AG investigation has been underway for 21 months, without any public announcement.

PRODUCTS and END MARKETS
A. CA has three major product areas:
-Large Computer monitoring (~50% of revenue): CA remains the dominant player in large computer monitoring (i.e. help desk software, asset management, provisioning) and competes with BMC, IBM (Tivoli) and HPQ (Openview); this market is growing roughly at 5-7% annually, and there should be accelerated growth over the next 1-2 years, as we enter a new mainframe cycle (could be 20%; IBM introduced the T-Rex in Fall 2003). CA has also been aggressively targeting the distributed computer monitoring market which has a faster growth rate and has been working on developing its channels to target this market.
-Storage retrieval (~20% of revenue): Storage is CA’s second largest product area, and CA has been challenged over the past couple of years, mainly by Veritas but also by a more aggressive EMC in software (bought Legato). CA has shifted its focus from the high-end of storage to the low-end and is in the process of building out a channel (hence all the CA, WSJ ads). Storage has been a declining area for CA over the past two years, but given the secular focus on storage as well as strong R&D marketing focus, CA should be able to turn the business around and grow in the low-end of the market. Furthermore, CA released version 9 of its storage software about 6 months ago and according to customers it corrects many of the flaws of the previous versions that had caused growth to stall (this could be a 10% growth business).
-Security (~15% of revenue): Security is CA’s fastest growing segment and grew 21% on annualized bookings level in F2003. CA has products such as intrusion detection, anti-virus and access control (i.e. authentication). CA largely competes with Symantec, Network Associates and IBM. CA’s approach for security differs from SYMC and NET as CA is coming from the high-end of computing and migrating downwards, while SYMC/NET are coming from the consumer desktop and trying to migrate upwards (CA has very small market share in anti-virus). Last month CA announced a partnership with MSFT to provide free antivirus software in order to improve CA’s brand recognition. According to contacts the campaign was successful as there were 500K downloads within the first day. Security should be a strong growth (20%+) area for CA for the foreseeable future as a result of both industry growth as well as CA share gains.

B. Unification of all of CA’s products: On-Demand Computing:
CA has outlined a strategy to lead corporations to ‘on-demand’ computing, a nascent, but often highlighted potential evolution of enterprise computing. In a nutshell, today’s severs are largely application specific, i.e. a corporation has dedicated servers for specific apps such as HR apps, financial apps, trading etc.. As a result of the growth in multiple applications, corporations have many dedicated servers, many of which are running at very underutilized levels. On-demand computing, similar to an electric utility, eliminates dedicated servers for specific applications, and instead allocates computing power dynamically, based on the needs of a corporation/application at a specific moment. CA, as one of the most diverse software companies, has been the most vociferous of driving towards an ‘on-demand’ environment. Any shift towards ‘on-demand’ represents a strong growth opportunity for CA.

THESIS
In my view, CA is a strong company with great free cash flow, and solid balance sheet, that is undervalued, due to its historic customer practices, the investigation overhang and a business model change that has artificially depressed both revenue and earnings for the past couple of years. I believe that all 3 of these factors will recede over the next two years and CA’s multiple/earnings will more closely resemble other large software companies. If one values CA based on its free cash flow of $2.20/share over the next twelve months, and use a multiple that is a slight premium to its bottom line growth rate (growth rate of 12-15%, using a multiple of 17x, which is a discount to its comps including MSFT) one derives a $37 stock price. As the overhang recedes, which I believe it will over the next 6 months (with a possible near-term catalyst in the next month), I think the stock will gravitate to the low to mid 30’s (it was at $29+ in early October). Furthermore, given solid software fundamentals and relatively low expectations, CA should not have a problem of meeting estimates over the next year, which in my view are pretty conservative. Finally, I believe that CA, through its ‘on-demand’ computing initiative and diverse product platform, has an opportunity to meaningfully differentiate itself from other software companies and show accelerated growth over a 3-4 year time period.

BULL POINTS
Changing Factor #1 – Customer Perception
Customer Perception – Then and Now:
I have spoken with a number of customers, software resellers and consultants to try to gauge the changing perception of CA in the marketplace, as I believe this change could represent the most important growth driver for CA over the next 2-3 years. According to customers, marketplace perception of CA in the late 1990’s was very poor. CA would sign five year deals, try to raise prices after three years and would be very difficult to negotiate with. Furthermore, CA would buy its competitors and discontinue development/support thus forcing customers onto CA’s platform (‘buy and die strategy’ – CA did over 30 deals). Customers began to notice a cultural change in early 2001, immediately following the model change to subscription based revenue and promotion of Sanjay Kumar to CEO. Clear changes include, much shorter deal length (average is under three years now), friendlier negotiations, minimal acquisitions, a new customer support call-center (650 people) and salesforce comp which is very dependent on customer satisfaction. Many of the resellers and customers that I spoke with said that CA has made huge strides, particularly over the past 6 months in improving its image. While there are some deals that CA will not get invited to, due to its past perception, my conversations indicate to me that those deals are becoming fewer and we are nearing an inflection point on CA’s market perception. As an example Gartner recently shifted CA from a ‘bottom-3’ company in terms of marketplace perception to a ‘middle of pack’ company. The change in customer perception, while there is still lots of room of left for improvement, was a consistent message that I received amongst all the constituents that I spoke with.

Changing Factor #2 – the Investigation
1. Internal board investigation – CA’s new board, launched an internal audit, headed by Bill Schuetze, a new board member and the former chief accountant for the SEC. In October the board announced that it had found questionable recognition practices from the March 2000 period. CA’s stock fell about 15% on that news. The CFO and three other finance people were forced to resign. Through a network of contacts, I have been able to learn a bit more about the issues. My understanding is that CA was recognizing revenue in March 2000 on deals that were signed in the last minute of a quarter, without requisite senior management (at customer) sign-off (the senior management sign-off would come a few days later-gray area). The revenue recognition issues are limited to questionable timing issues and do not impact the absolute revenue number, nor cash flow. Furthermore, these practices were completely stopped when CA shifted its business model in October 2000. CA’s CFO was let go, as he was CFO at the time and any hint of questionable practices is being dealt with very strongly. My understanding is that the board is close to wrapping up its investigation and there could be an announcement over the next few weeks (although I am not sure about that). It seems to me that the worst that can happen at this point would be a revenue restatement for periods more than three years ago and at an extreme, the CEO Sanjay Kumar, would be asked to step down. The CEO leaving might be a positive, as he is the last hold-out from the old regime.

2. The SEC/US AG investigation – the SEC and US AG (Eastern District of NY) has been investigating CA for the past 21 months (Feb 20, 2002). The investigation is focused on revenue recognition policies based on CA’s old business model. Based on looking through press releases and talking with sell-side analysts, it seems as if the SEC and US AG have not commented nor issued preliminary findings over the course of the 22 month period. As a result, I would conclude that the investigation has not turned up anything significant and is probably somewhat stale at this point. Given the length of time that the SEC has used to this point, I do not expect significant material findings in that investigation.

Changing Factor #3 – Business Model
CA’s business model change has depressed revenue and earnings and therefore multiple valuation off of earnings is largely irrelevant. In October 2000, CA introduced its ‘FlexSelect’ licensing program, which represented a radical change in the way CA interacted with customers as well as booked its revenue. Under the new program, CA began to defer nearly all of its bookings on to the balance sheet, recognizing the revenue on the income statement ratably over the length of the contract. Previously, CA had recognized license fees upfront, and only deferred the maintenance portion of the contract. As a result of the change (as well as the downturn), CA’s revenue fell by nearly 45% from F2001 (march 2001) to F2002 (march 2002). Similarly pre-tax earnings fell sharply (CA did not alter expense recognition). However, the drop in free cash flow from ops was much smaller, only 10% y/y, as the business model change had limited impact on free cash flow. Today the key metrics to use to judge CA’s business are ‘annualized deferred subscription revenue’ and cash flows. The annualized DSR metric reflects renewals from the old business model, i.e. contracts that were 4-10 years long, new business model renewals (minimal at this point) and new business. CA has been growing its ADSR at a mid-teens rate for the past 5 quarters, although it slipped to 7% y/y in the September quarter. The second sell-off in CA’s stock in October, came largely in response to the 7% bookings growth (analysts had expected 9-10%). I believe that these numbers are lumpy, include many different factors and would not take a one quarter drop-off as an indication of deteriorating company fundamentals. Importantly, in 2005, CA’s revenue should become more aligned with its expenses as the majority of old contracts (5 years in length-recognized in C2000) will be in renewal phase.

Customers that I spoke with cited CA’s ‘Flexselect’ licensing program as the most customer friendly software licensing program in the industry. CA’s renewal rate has remained solid at 80%, per its 10K. Aside from improving customer perception, the subscription model also enables significantly better revenue visibility, as most of CA’s quarterly revenue is from contracts booked in prior quarters. This minimizes the risk of missed quarters and last minute deal signings and greatly improves linearity, something absent in many enterprise software companies.

COMPANY SPECIFIC RISKS:
-Headline risk on the investigation front: I have tried to dig around amongst people that are close to the company to gauge the internal feeling on the investigation. The feedback from talking to employees was generally that this is a changed company and they are not concerned about revenue timing issues from 4 years ago. While risk remains, I think the key points that I highlight above, length of SEC investigation, board findings of only revenue timing issues, not cash flows, gives me comfort that we have seen/heard the worst and the next step is resolution.

-Lumpiness of bookings: Since CA’s revenue and earnings are much more linear and predictable at this point, the main quarterly risk is bookings lumpiness. CA reports a bookings number as well as deferred revenue on its balance sheet. In its last quarter its booking were slightly below Street forecasts and stock fell 7-10% following the report. Based on customer conversations, I believe that CA is no longer managing for end of quarter and large deals can fall in and out of quarters, creating an element of lumpiness. These issues are short-term in nature.

CORPORATE GOVERNANCE OVERHAUL:
CA, for the past three years, has been very focused on being the corporate governance model citizen in tech-land. As an example, CA is one the first tech companies to expense options (forward EPS estimates include options). More importantly, CA has done a large scale board overhaul since 2000. CA currently has 10 board members and only two insiders. The board members include the SEC’s former Chief Accountant of the Division of Enforcement, a Harvard Business School professor that is a well-known thought leader on boards of directors/corporate governance, and a former U.S. Senator.

Sanjay Kumar:
In August of 2000, Sanjay Kumar was appointed president and CEO. He became Chairman in November 2002. Many on Wall Street have a strong dislike for Sanjay particularly following the very large option grant ($1B) given to the top 3-5 employees of CA in 2000. In contrast, the employees and consultant that I spoke with were much more positive and said that he is a great leader and has been able to bring massive strategy/cultural change to the company. Two recent examples of Sanjay’s changes are: 1. Customer Research Organization which is a 650 person business unit that is in charge of interacting with customers to help improve technical support, product development and enhance customer relationships and satisfaction levels. 2. Acquisitions – After being very acquisitive in the 80’s and 90’s, CA has stopped all major acquisitions and are more focused on internal development and organic growth. Both these changes, according to contacts are very significant and represent major accomplishments for Sanjay.

Subscription model:
CA shifted its revenue recognition model from upfront licensing fee recognition to pro-rate subscription recognition. See below for an example:
$300 contract for 3 years:
Old Model
$200 License Fee recognized in year 1
$20 Maintenance Fee in year 1
$20 Maintenance Fee in year 2
$20 Maintenance Fee in year 3
$40 Financing Fee over like of contract


New Model
$100 License Fee in year 1
$100 License Fee in year 2
$100 License Fee in year 3
As a result of the change, as well as the long nature of contracts under the old business model, revenue and earnings have been artificially depressed since October 2000 and through the first half 2005 (initial contracts under the new model were 4-5 years in length). These changes did not impact expense recognition and did not have a significant impact on cash flows. The benefits of the subscription model include: better visibility, improved linearity, improved customer satisfaction (shorter contracts, more flexibility).

Illustrating the cascade effect of the subscription model:



VALUATION:
I believe the best way to value CA is off of its cash flows, as earnings are artificially depressed. My free cash flow estimate for the next twelve months is $2.20 per share. Assuming a 17x multiple, which is a slight premium to my assumption of 12-15% bottom line growth rate and a meaningful discount to comps (25x-30x), I derive a $37 fair value. CA currently has $2.3B in debt and $1B in cash. The debt is comprised of senior notes and two convertible notes. CA also has $870M of an unused credit facility.

Catalyst

Catch up from change in accounting policy
Resolution of Accounting Investigations