"All models are wrong, but some are useful" - George Box
I've always been skeptical of financial models. First, personal experience has taught me that, even with access to what essentially amounts to "insider information" and deep industry expertise, I still find it impossible to reliably predict where my own businesses are headed. Second, when predicting more than a year or two out, any errors and invalid assumptions will be compounded.
It seems the height of insanity to say "based on an extremely optimistic estimate of 2021 earnings, we think this company trades at a fair value given a 35x forward multiple", but professional analysts do it all the time (especially for fashionable growth stocks). Perfection for the next few years is essentially priced in, and any negative deviation, however slight, will likely result in a catastrophic drop in stock price.
That's not to say that models are useless. You just need to take them in context, be aware of the limitations, and make sure that you are confident in your larger assumptions before fretting about the details.
Precision vs Accuracy
Nokia (NOK) was one of the first companies that I tried to financially model. The models are fairly simplistic, but that was the intention. My goal was to try to get the "big picture" assumptions right, rather than get mired in too much detail. In fact, since I am not an expert on the telecom industry, I felt that these were the only types of assumptions I was capable of making with any chance of accuracy.
I feel like this approach may lead to lower precision, but higher accuracy, a difference which is illustrated by the following images:
Furthermore, I tried to keep all of my assumptions conservative to provide a larger margin of safety.
Big Picture Assumptions
First, a quick summary of Nokia in 2019. Most people probably know them for their old mobile phone business, which got crushed by iPhone and Android smartphones. Since then, they've reinvented themselves as a networking and telecommunications equipment company (and are one of the largest 5G equipment suppliers in the world). They also have a licensing business for their wide portfolio of patents and the Nokia phone brand (which is now controlled by a separate company). Remarkably, mainstream financial news sites still incorrectly attribute NOK stock moves to their consumer phone launches, showing how widely misunderstood this company still is.
The conservative case for NOK assumes the following (note that these were written in summer of 2018, and some of these points may be already out of date, but I wanted to accurately capture my thinking at the time, without hindsight bias):
- Revenue will decline by 6% YoY in 2018, even though the Q1 2018 earnings report says that “Nokia sees market conditions improving and 5G accelerating further, with strong momentum building by year end. Nokia now sees a stronger primary addressable market for its Networks business in full year 2018 and expects its Networks business to outperform its primary addressable market in full year 2018”, which would imply an increase. Many of these contracts are already “in the bag.”
- After bottoming out in 2018, future years will only see revenue increase by the rate of inflation (2%), despite the fact that 2018-2019 will likely be the bottom of the current wireless infrastructure spending cycle, and we are at the beginning of a large scale up in 5G spending.
- NOK will only maintain (not grow) market share in global telecom equipment spend vs Ericsson, Huawei, and ZTE. Despite the facts that ZTE might go bankrupt (even if they don’t, the uncertainty regarding their near term prospects will severely hurt their ability to win bids for a while); Huawei will not likely get access to the US and other western markets due to security and political concerns; and the CEO is aggressively hinting at increased market share in nearly all segments.
- The diversification away from pure telecom customers to Webscale and Enterprise customers will not materially increase revenue. Despite the fact that it currently accounts for 5% of revenue, is a large untapped market opportunity, and has recently growing at a rate of 20% YoY (which would imply that this opportunity alone could account for an additional 1%+ YoY total revenue lift)
- Operating margins will be at the absolute lowest end of guidance. Despite the fact that they’ve been successfully winning business and cutting costs over the last 2 years, laying people off, and aggressively restructuring as a result of the Alcatel-Lucent merger. And despite the fact that the current CEO, Rajeev Suri, was responsible for cost-cutting and turning the Nokia Siemens Networks business profitable as CEO of the division starting in 2009.
- Licensing revenue growth will decrease by 20% YoY immediately after 2020. This is despite the fact that most of the revenue is recurring, and that management has guided for a 10% CAGR from 2018-2020. This is also despite the facts that NOK has 20,000+ patents, is filing an additional 1,300 patents per year, was a key player in developing the 5G spec, and owns one of the most storied R&D centers in US history (Nokia Bell Labs)
The growth assumptions seem reasonable, but the margins (even on the lower end) depend on the guidance of the management team. Ultimately, I had to make a judgement call based on the trustworthiness and incentives of the management team.
To assess trustworthiness, I found a particularly interesting line about the CEO in Transforming Nokia: The Power of Paranoid Optimism to Lead Through Colossal Change, a recently published book written by the Chairman of Nokia:
Rajeev is always extremely polite and rather soft-spoken. But despite being somewhat introverted, Rajeev reaches out to people. As a salesperson, he was known for trying to keep his word to customers under all circumstances, and he gained a lot of trust from teleoperators because of that.
It seems reasonable to assume that someone who was adamant about keeping his word to customers would also try to keep his word to investors. But, talk is cheap, what about actions and incentives?
Starting in 2018, Nokia introduced a "co-investment" plan for the CEO and Executive Team to participate in:
Under the co-investment arrangement, the participants will be offered a matching award of two 2018 Performance Shares for each Nokia share that they purchase voluntarily with their own funds from the open market, with the payout of the Performance Shares subject to actual performance. For each participant, the arrangement is offered in addition to their normal annual long-term incentive award, and the maximum investment value corresponds to their normal annual long-term incentive award set by the company.
Making some rough assumptions, it seems like nearly every member of the executive team has recently maxed out their voluntary co-investment purchases, with roughly €10MM+ of personal funds at stake.
The cynic could argue that since they are receiving 2 free shares for every one share they purchase, shares could essentially lose 2/3rd's of their value and they'd still break even - but the shares are "subject to performance", and it is unlikely that they will be paid out if performance severely misses guidance, so I think they have some real skin in the game.
Back to the Model
With all those assumptions out of the way, here are the various models I put together for Nokia:
I tried to tackle valuation using a few different methods: Discounted Cash Flow, Operating Profit, and Bottom Line (which you can access via the different spreadsheet tabs on the bottom). Nokia also has a lot of (supposedly) one-off charges associated with the merger / restructuring which I tried to back out to get a clearer picture of "true" earnings potential.
Please note that these models were last updated ~6 months ago, so some of the numbers may be out of date, but again I wanted to accurately capture my thinking at the time of analysis. This was also the first time I attempted to do a DCF analysis, so it's also highly possible that I did it completely incorrectly (feedback is welcome).
With that said, I think the most important takeaway for me during this process was not in the precision of the numbers themselves (since the exact numbers are likely wrong), but in whether or not I believed they were generally accurate. Am I in the right ballpark? Am I on target? Is the model useful in making some general assumptions about value?
I think the answers to these questions depend more on the high level assumptions (see above) than on the minutiae and exact formulas. Importantly, many of those assumptions rested on subjective (and arguably common sense) analysis rather than quantitative metrics.
If you buy into these assumptions, and believe that they are conservative enough to give you a sufficient margin of safety, then I think I safe price to pay for Nokia is around $5.60.
Unfortunately, NOK has had a recent run-up of late, but there were multiple opportunities to buy at this price over the past few months...
...and there may very well be good opportunities to buy in the future (assuming the fundamental story doesn't change).
Disclosure: I am long NOK.