Investment update on Valero #1
Updating DCF valuation based on Crack 3-2-1 pricing for VLO
Dear readers, apologies for the late release of this week’s post. I was unfortunately stuck in bed for a few days with the flu and that resulted in a bit of a delay. This week, I include a fairly short post length wise although it took quite a bit of time to put together offline due to the time it took to parse the historical data and the underlying analysis.
I wanted to spend time talking about tackling prediction and valuation in DCF analysis. Specifically, what do we do in cases where we do not have a nice steady growth rate in FCF that is the textbook 2 stage DCF analysis. I use Valero (VLO) as a case study parsing historical data going back to 2010 to build up the problem statement. I then conclude the post with an updated valuation of Valero based on the improved valuation models and summarize some interesting implications from this workup as well.
I hope you find it informative.
An overview of the problem statement
In our original Investment thesis for Valero, we built a simple DCF valuation model. However, we also recognized that the refinery business actually depends heavily on the cost of raw materials. As I originally wrote,
Valero can generate a ton of cashflow but their business is quite cyclical depending on the inputs (crude oil price) and demand for the refined products. The difference between these is what is referred to as the crack spread which can vary considerably leading to large fluctuations year to year in free cash flow (FCF).
In other words, how much money the oil refinery business makes in a given year isn’t entirely up to them since it depends on how much the raw products cost and how much they can sell the finished products for. We call this the 3-2-1 crack spread. As a reminder, the 3-2-1 crack spread assumes using 3 barrels of crude oil (input) to produce 2 barrels of gasoline and 1 barrel of diesel (output). The spread is the incremental value that the finished products can sell for higher than the crude oil per barrel.
At the time, I used an approximate estimate for anticipated 3-2-1 crack spread for FY2025 to estimate the FCF then assumed a constant 4% growth rate in this FCF rate over time. However, if we look at historical VLO FCF since 2010, this constant growth rate of 4% is clearly not reflective of reality. In fact, FCF was pretty stable for a number of years in 2015-2019, dropped negative in 2020, peaked in 2022, and then declined in 2023 and 2024.
In this post, we attempt to take a more sophisticated view of modeling FCF by parsing historical data.
Valero’s Free Cash Flow does depend on 3-2-1 Crack spread.
As we have hinted at in this post, Valero’s FCF depends heavily on 3-2-1 crack spread. A plot of historical annual FCF vs average 3-2-1 crack spread shows a strong association between the two parameters. Closer inspection of the data will also show that later years tend to be slightly above the regression line indicating a gradual improvement in either total capacity or efficacy for the business.
Interesting Aside: Valero’s Free Cash Flow does not depend on crude oil price.
Usually, when deal with oil companies, the first thing that investors monitor is the crude oil price. Not only does it impact raw input prices, it also impacts how much oil companies are willing to spend on exploration, drilling, etc. When crude oil futures “CL=F” increase, oil stocks tend to also increase in price. That’s why it’s particularly interesting that for Valero, I found their FCF to be negligibly impacted by crude oil prices.
A review of the historical data shows that VLO FCF vs average crude oil price are only minimally associated. This makes sense to an extent since oil refineries can pass on much of the higher material costs downstream so long as demand is robust.
Bottom line, we should be actively monitoring the 3-2-1 crack spread far more than the crude oil price when thinking about Valero’s annual performance.
This outlines roughly how we should build a DCF model to account for asset pricing based on 3-2-1 crack spread instead of assuming a constant growth. The paid subscriber portion of this article reviews how to derive 3-2-1 crack spread, the historical 3-2-1 crack spread performance, and develops the actual model to update our valuation of Valero.