The Dark Arts - Fair Value Adjustments

My last post explored my attempts to bend ChatGPT to my will and create an analysis tool to rule all others - a Sauron-esque, all-seeing eye, penetrating distance and detail to be the undisputed master of salmon farming analysis. If you read the post, you’ll know that my efforts were laughably distant from this kind of glory, and I hit a wall of confusion and frustration.

One thing, however, got me thinking about my approach to reading financial reports and presentations. I had asked ChatGPT to extract all the data from the interim financial statements and accompanying financial notes and ignore the summaries, management commentaries and key figures sections. My reasoning was that the formatting of the information was more consistently represented, more granular, more bound by accounting rules and less impacted by commentary intended to shape the conclusions the authors want you to make.

While I am an avid reader of financial reports, I have a few complaints: 1) they are fundamentally backward looking 2) key figures like mortality, average harvest weights, feed conversion ratios etc. can be presented in ways that can obscure important trends and 3) The commentary can, at times, give the wrong impression to a reader who doesn’t go deeper than the first few pages. One of the outcomes from my ChatGPT fiasco was the realization that I wasn’t following my own advice and had not been looking deeply enough at the most important information at the back of the document. In trying to predict the future, the best information is often deep in the document.

Fair Value Adjustments

Fair Value adjustments are an accounting practice designed to smooth out earnings in a volatile market environment. In salmon farming, these adjustments are designed to apply a consistent value to living inventory. For example, if you have a population of fish that are 3 kgs today, and you expect to harvest them in Q1 2025, you would model your expected growth and costs between today and when they will be harvested, apply an expected market price in Q1 2025 and then declare the difference between your expected revenue less production, harvest and sales costs to arrive at an adjusted value of the biomass based on future earnings. The Q3 2024 earnings report for Kaldvik contains an excellent detailed explanation of the philosophy behind the adjustment on pages 21 & 22.

The Fair Value adjustment is then reflected as non-cash revenue on the income statement and will inflate or deflate the declared EBIT for the company. I have long been skeptical of these adjustments and tried to ignore them as much as possible. Two reasons for this; I am not an accountant and the whole concept seemed bizarre, and in my early days of reading financial reports, I recall the commentary of a Chilean company that was losing millions due to disastrous production costs and yet, miraculously, posted a profit thanks to an enormous fair value adjustment. An experienced reader would have seen through the commentary immediately, but I was young, dumb and had been taking their commentary at face value.

One of the few benefits of aging horribly is the opportunity to reflect on pre-conceived notions and see the value in concepts I had previously ignored. I’m still confused by the adjustment process. It can be highly subjective and varies tremendously from company to company and, I think, is open to considerable discretion in terms of how much or how little supporting information is presented. Fair value adjustments and, more importantly, the notes explaining them can be a source of valuable forward-looking information and a salmon farming company and their view of coming quarters. As an investor or analyst, I will draw confidence from companies who provide some granularity on their Fair Value processes.

Price and quality information:

Part of the FV calculation rests on price expectations for the current living inventory. With a production cycle of not less than a year, the calculation relies on a price prediction by quarter. Not all companies provide this level of detail, but a few do.

Bakkafrost Forward Price Estimates - Q3 2024

Masoval Forward Price Estimates - Q3 2024

The European producers who share their forward-looking price estimates seem to be fairly consistent in pulling their estimates from Fish Pool. I’ve only looked at one Chilean operator in writing this post (they didn’t provide their estimates) but it would be very interesting to see how their projections compare.

The Q3 2025 price forecasts have the average price landing uncomfortably close to production cost and a mid-summer operational wobble could push producers into the red.  Given the fallout in share prices for companies like Grieg and Masoval who reported losses, I would expect Q2 2025 to have the maximum harvesting of any farm even remotely close to harvest size and an absolute minimum harvest level in Q3 if you can manage it. If you buy salmon company shares, a potential arbitrage would be to buy a few days after a weak (in relative terms) Q3 report is issued and sell just after Q2 when the company is likely to have posted strong Q1 and Q2 earnings. Given the solid long-term prospects for most salmon companies, that kind of play is likely extremely flawed.

Biomass status

Below are two examples.

Grieg Price Related Assumptions Q3 2024

The first is from Grieg Seafood Q3 2024. The premium price in the quarter looks optimistic compared to prices reported in media (which had premium Norwegian fish selling for much of the quarter in the 70 – 75 NOK/kg range). The average superior share in Norway is low but seems optimistic based on what we have been hearing about sea lice and jellyfish over the past few months and compared to the results for 2023. The 20% discount on non-superior graded fish in Q3 – 2024 is more than 3 times higher than the same period in 2023 and a lot higher than the result for FY 2023. I don’t know if this indicates a significant deterioration in biomass status, a market saturated with downgraded fish or a conservative prediction. Assuming they have consistently provided this information in previous reports and will do so going forward, this is something to keep an eye on.

Bakkafrost Incident Based Mortality Q3 2024

The second example is from Bakkafrost. I’ll do a deeper dive on incident-based mortality in a future post, but the snip shown is from their FV adjustment section, so I’ll touch on it here. The main takeaway is how much their Scottish operations have improved compared to 2023. Costs in Q3 2024 are 1/6th of those registered for exceptional mortality compared to the same period the previous year. Year-to-date extraordinary mortality for this year is 1/3rd of 2023.

When evaluating the performance of a salmon farming company, the trend in the data is very important. In the Grieg example, the arrows point towards deterioration while the Bakkafrost example is pointing towards an improvement. Taken alone, these data points may not seem significant but should not be ignored.

Biomass Costs and Volumes

I’ve provided three examples of interesting information.

The first example is from Mowi. The snip shows stable biomass development and a positive adjustment in every region except Canada (which is worth thinking about) and a cost-to-stock of 4.74 Euro/kg. Not shown in the clip but available on their report are the costs from December 31, 2023, which work out to 4.91 Euro/kg. A decent trend on both volume and cost.

Mowi Cost to Stock Q3 2024

The second example is from Grieg. Looking at the number of fish and the biomass, they have considerably fewer harvest-sized fish on the books compared to the same period in 2023 – 4,000 tons less which will likely translate to a revenue drop in Q4 and Q1 2025 of somewhere close to 30m Euros at an average price of 7 Euros per kg.  (Don’t @me on my math – I’m just trying to make the point that it’s a big miss)

The production costs have risen slightly, from 64.8 NOK/kg in 2023 to 68.8 NOK/kg in 2024. Given how poor their results were for the quarter, I’d be inclined to see how much extraordinary mortality write-offs have impacted these numbers.

Grieg Cost to Stock Q3 2024

The third example is from Masoval. I like this example for a couple of reasons. First, by giving the biomass, count and cost, it is possible to build a very primitive unit cost model. For example, salmon less than 1kg have an average weight of 592 grams, a cost per fish of 62 NOK and a cost per kg of 105 NOK. Looking at the same period in 2023, the fish in this category had an average weight of 284 grams, a cost per fish of 38.1 NOK and a cost per kg of 134.1 NOK. So nice trends on average weight and cost per fish but perhaps a negative trend on cost.

Masoval Cost to Stock Q3 2024

Conclusions?

OK - pretty deep into the weeds on this one. In the unlikely case that you are still reading, I want to give you a few take aways:

1) You are deeply nerdy and should, perhaps, seek professional help.

2) Near-term success in salmon farming far more dependent on a company’s living biomass than it is on previous performance and you have to dig for key forward-looking information. Commentary and summary narratives put the best spin on things and reflect the natural optimism of farmers. If you are assessing risk, you want a pessimistic take.

3) Real insights can be gained by reviewing period-over-period trends in the data and how it is presented. Given the long production cycle for salmon, trends are everything.

Thanks again for reading. Feedback can provided in the comment sections below, via my LinkedIn page, or by email at Info@AlanWCook.com