I've been misallocating my most valuable asset: Time
Because summer is meant for beer, not repairing financial data
Hello again, and thanks for reading. This month I am making some fairly significant changes after recent epiphanies — triggered by the odd combination of the Finnish summer sunshine and an offhand comment made by a friend of mine.
Falling victim to the Sunk Cost Fallacy
Despite my firm belief in the systematic investment approach, I still find myself affected by the various psychological, behavioural elements far more often than I want to admit. In particular, I’ve (finally?) come to realise that I have fallen victim to the Sunk Cost Fallacy.
“The phenomenon whereby a person is reluctant to abandon a strategy or course of action because they have invested heavily in it, even when it is clear that abandonment would be more beneficial” — Oxford Languages
For several years now I have been harvesting, cleaning and compiling a kind of financial “data lake” — a collection of different financial data pulled from various sources (some free, some paid) and used as input into my backtesting/investment engine software. As mentioned in previous posts, I have spent quite literally thousands of hours on this, and it is a constant battle to keep the data flowing, on time and without data corruption, data feed interruptions, outtages, and so on — especially when scraping data or using free sources whereby sadly, the quality of the outcome often matches the price paid.
Discussing my ongoing data woes with a friend recently, I mentioned that I had found a paid data provider which (aside from a few data points missing) seems to offer 95% of what my existing complicated, fragile infrastructure achieves, whilst offering the service as a single, convenient data feed service at an affordable cost. When I said that I was tempted to swap over from my existing setup to use this new alternative instead, but that I “am nervous about abandoning all the work I had done” and “could I trust this new source instead, what if it goes away, etc etc, maybe I should keep my existing system and run both in parallel”, his reply was the deliciously obvious slap that I needed:
“You know, you DO do an awful lot of work,
with the principal aim of not needing to work“
For roughly 15 years, he and I have discussed investments and trying to achieve financial freedom — reaching so-called “Fuck You” money, whereby someone has enough wealth and passive income to be able to quit any bullshit job or situation without fear of repercussion or financial stress. Instead, with Fuck You money, one can instead pick and choose a life of more meaningful, appealing work. Financial freedom to do what is interesting and what matters.
My friend’s point though, was that I was spending ALL my free time in the pursuit of gaining more free time, a bit like the story of the finance guy and the fisherman.
I was reluctant to abandon my existing code and structures based NOT on the value it was delivering to me, but on the simple, stupid fact that I had spent a lot of time to implement it — the very embodiment of Sunk Cost Fallacy.
Enjoying time while the sun shines
This was made doubly obvious by the fact that Summer — in the form of glorious warm sunshine — has finally arrived in Finland after what felt like the longest, dreariest winter in the 8 years I have lived here.
Staring at my computer screen and trying to repair yet another flaky data import so that I can generate my monthly performance reports, all while the sun shines outside my window and summer beer terraces call my name, suddenly seems like the absolute height of stupidity.
Especially when there is what seems to be an easier, more efficient, more effective solution so readily available. So with that decided, I will make the change and adopt this new data source going forwards (or, as with our investments, until something better comes along).
So what am I saying?
Nothing will change about the Two Crows approach — I will still product monthly model long-only stock portfolios for the 8 currencies and report the results each month. The selection process and the goal (to target a 15% annual return over multi-year periods, doubling our portfolio size roughly every 5 years) remains the same.
However, I probably need another week or so to fully integrate this new data source with my investment software. Unfortunately, it seems that some of our earlier stock picks are NOT available in the new datafeed, meaning those stocks might not be eligible for inclusion in the model portfolios going forward. I am in contact with the data provider to see if they can include the missing items, but in the case that they cannot, I will restate past portfolios to remove stocks such as:
DKK - Novozymes / Novonesis
EUR - Comer Industries
EUR - United Bankers
INR - Innovana ThinkLabs
JPY - Starts Publishing Corporation
SEK - LumenRadio AB
If cases where a stock must be excluded, our model will pretend that we held cash instead of buying that particular stock (i.e. I will not redistribute that cash across other holdings for past dates). This may in some cases result in improved performance (if the missing stock has been declining in share price since purchase) or detract from performance (in those cases where shares have increased in price since purchase), but I expect the net effect to be fractions of a percent in either direction.
I will post an update when the data feed swap is in place, and once I have confirmation about the currently-missing stock items. In the meantime, thanks for your patience.
Love the honesty!