Crunching numbers on Alberta oil
Greetings patrons,
Happy New Year. I hope you are safe and healthy during these uncertain times. In this research update, three things:
- Analyzing oil-and-gas depletion in Alberta
- Preferential attachment of vaccines
- Computer obsolescence
1. Analyzing oil-and-gas depletion in Alberta
For the last few months, I’ve been working with researcher and activist Regan Boychuk on an exciting new project. Boychuk studies the (bad) behavior of the oil-and-gas industry in Alberta, Canada. Back in September, he asked me to join his project. Since then, I’ve been crunching the numbers on Alberta’s oil-and-gas production.
Boychuk’s goal is to get the oil industry to clean up its mess. For decades, Alberta oil companies have been drilling wells, sucking out the oil, and then abandoning their wells. The result is that hundreds of thousands of derelict wells dot the Alberta landscape. Boychuck wants to change that. His dream is to force oil companies to clean up their mess.
The catch is that the clean-up liability is so massive that it would bankrupt the Alberta oil industry. And that’s the point. Boychuk wants to simultaneously clean old well sites, kill the Alberta oil industry, and fund a transition to a sustainable economy. Given the state of Albertan politics, this dream isn’t going to be fulfilled any time soon. Still, it’s worthwhile to think about.
To get started on this project, Boychuck asked me to estimate historical production curves for every oil and gas well in Alberta. (Boychuck got his hands on some very expensive and very proprietary data that he kindly shared with me.) After much head scratching and many lines of code, I’ve finally crunched the numbers. I’ll publish the full results in early January in a post I’m tentatively calling ‘A case study in oil-and-gas depletion’.
Here, though, is a teaser chart for your enjoyment. The figure below shows aggregate oil-and-gas production in Alberta, resolved to the individual well. Color indicates the year in which the well was drilled. The resulting ‘stratification’ shows you what oil-and-gas production would have looked like if drilling had stopped in the given year. You can see that as time goes by, new wells are depleting more rapidly.
Historical production of oil and gas in Alberta, resolved to individual wells.
When I write up my results, I’ll go into more detail about what’s going on here. But for now, let me give you a sense for how much data you’re seeing in the figure above. I’ve plotted monthly production data for over 300,000 wells. That amounts to about 50 million data points.
A computing side note: It turns out that most graphing programs don’t like to plot 50 million points of data. I usually use ggplot2 to make my charts … but it falls apart when fed millions of data points. So to make the chart above, I had to learn how to use gnuplot. It’s not as convenient as ggplot, but it’s blazing fast. In the future, I’ll be using gnuplot anytime I have to plot a big dataset.
Stay tuned for my write up about oil-and-gas depletion in Alberta.
2. Preferential attachment of vaccines
Neoclassical economists talk a lot about ‘market failures’ — instances when markets have failed to distribute commodities in a way that is ’efficient’. Many heterodox economists have adopted the same language when highlighting problems of resource distribution. I think that is a mistake.
The distribution of COVID vaccines is a prime example. The roll out of vaccines has been horribly unjust, with the vast majority going to rich countries. It’s tempting to call this inequity a ‘market failure’. But doing so assumes that if markets worked ‘properly’, they would distribute resources equitably. That, however, is not what markets do. Markets distribute resources according to purchasing power. In other words, the rich get what they want and the poor suffer what they must.
When framed this way, the distribution of COVID vaccines has been a triumphant market success. Across the world, vaccines have chased purchasing power. The result is that vaccination rates correlate strongly with GDP per capita.
It recently occurred to me that there’s a simple way to model this market ‘success’. It’s called ‘preferential attachment’ — a model originally proposed to explain the distribution of connections within networks. To apply the preferential-attachment model to vaccines, we imagine adding vaccines to the world one by one. The catch is that the probability of getting a vaccine is proportional to purchasing power (i.e. income per capita). The result will be a wildly inequitable distribution of vaccines — much like what we see in the real world.
While this distribution is undoubtedly a moral failure, it is also an undeniable market ‘success’ — an outcome of the one-dollar-one-vote ethos. In an upcoming post, I’ll explain the details of this model and ponder what the future distribution of vaccines might look like.
3. Computer obsolescence
Continuing on the topic of market ‘successes’, permit me a short rant about computers. Computer manufacturers are, for the most part, in the business of selling new computers, not maintaining old ones. The result is a perverse market ‘success’, in which manufacturers earn profits by making computers that are designed to be defunct in a few years.
Here’s a case in point. Back in 2017, I bought an Android tablet that quickly became my daughter’s go-to media device. Over the years, though, software became more difficult to maintain. Often, I couldn’t find apps through the Play store … even though I knew these programs existed. The problem, I discovered, was that my tablet had not received a single OS update since I bought it. New tablets now come with Android 11. My Samsung tablet was stuck running Android 5.
The fact that my OS was so outdated confused me, as I had enabled ‘automatic updates’. But after some digging, I discovered that this update option is mostly a ruse. Android manufacturers typically push OS updates for a year or so, after which they abandon the device.
That’s annoying. But what’s infuriating is that these same manufactures lock their devices in order to prevent you from manually updating your device. In other words, they lock in obsolescence.
After learning about this practice, I was determined to give my tablet manufacturer (Samsung) a big middle finger. Instead of buying a new tablet, I learned how to install an updated (3rd party) version of Android. I ended up putting Lineage OS on my tablet. And in the process, I completely removed Google from the device. The results are fantastic. The new OS is snappy. And without the bloatware (that Google previously barred me from uninstalling), I reclaimed a significant amount of storage.
This tablet project got me thinking about computer obsolescence in general. Everyone knows that computers seem to get slower with age. The reason is that software programmers tend to rely on Moore’s law (the exponential growth of computing power) to solve their problems. If computer power was constant, developers would have to be careful about which new features they pushed to users.
Moore’s law solves this problem. Since computers have tended to become more powerful with time, developers tend to push every new feature that comes to mind. The result is software bloat: your computer is filled with software (often running in the background) that may be useful to some people, but is useless to you. As the bloat increases, older computers slow down.
Many people view this obsolescence as inevitable. But it is not. I first realized so back in 2011 when I was running Windows XP on (what was then) an old computer. The device had become a pain to use — taking a good 5 minutes to boot.
On a whim, I decided to install Linux on the machine to see what it would be like. I still remember my surprise. When I booted into Ubuntu (a flavor of Linux), it was liking being on a whole new computer. This is when I realized that computer obsolescence isn’t inevitable. And it’s when I vowed never to use Windows again.
I’ve now been a devoted Linux user for over a decade. But while Linux distributions largely avoid the bloat of Windows, they are not immune to feature inflation. And that is not necessarily a bad thing. For many users (who have new computers) having rich features makes their lives easier. But for someone like me who runs old computers, every new feature slows things down.
Annoyed by this trend, I recently decided to switch from a full-blown desktop environment (I was running Linux Mint MATE) to a minimal tiling window manager called i3. After the switch, I got the same feeling of elation that I had back in 2011 when I first switched form Windows to Linux. My old computer felt snappy again.
Of course, there is a trade off between speed and ease of use. (i3 has a steep learning curve.) That said, my experience goes to show that computer obsolescence is not inevitable.
Until next time
That’s it for this update. Thanks for supporting my work, and best wishes for 2022.
Cheers,
Blair