He uses statistics as a drunken man uses lamp-posts – for support rather than illumination.
Data is like coffee shops. These days it is everywhere and in abundance. It doesn’t matter whether you are building a hockey team, managing investments, or picking the fastest route home, there are a lot of numbers being crunched. We could be living through the first period in history where statisticians are cool.
New technology has afforded us the ability to access and process vast amounts of data. In theory, this should enable us to make more informed and rational choices instead of relying on ‘gut feel.’ But until algorithms are making all our decisions and driving our cars, we humans have to sort through and interpret this overwhelming flow of information.
The first challenge is selecting and determining what measurements are relevant. This can be especially tricky when confronted with divergent or contradictory evidence and is further compounded by our tendency to concentrate on material that confirms our pre-existing beliefs. Ultimately, what we choose to focus on and ignore will shape the views we form.
Once the data has been selected, it must then be analyzed. When interpreting the facts, we must be wary of any disconnect between what the numbers say and what people feel to be true. In such cases, evidence-based conclusions may not jive with the behaviour of consumers, voters, and market participants. As seasoned investors know, despite what the numbers indicate, markets can go from irrational to very irrational and stay there for a long time.
There is also the question of the quality of the data and the methods being used to analyze it. This is not just a shot at fake news, but an important consideration for scientific research. This why academics often review the findings of their peers, sometimes with surprising results. After all, as statistics show, 80% of statistics are made up.
Even when the data is good, we run the risk of confusing correlation and causation. A fun illustration of this was presented over two decades ago by David Leinweber and Dave Krider. They found a 75% correlation between butter production in Bangladesh and the S&P500. Adding US cheese production improved this to 95% and introducing a third variable; sheep population, resulted in a regression that explained 99% of the US stock market. As nonsensical as this seems, they apparently still get phone calls from people wondering what the butter indicator is saying.
Finally, there is the illusion of precision that can come from cold hard facts and numbers. There is a certain amount of confidence that is inspired by the exactness of 4.523. This can lead us to be precisely wrong rather than approximately right. With all the tools available to us, it is easy to overemphasize what we can measure and ignore what we can’t.
We have no doubt that making decisions based on hard evidence helps us avoid hidden biases, popular misconceptions, and wishful thinking. But a plethora of information also makes the decision-making process extremely complicated. Perhaps the greatest irony of the preponderance of data is that good judgment matters even more.
Aside from the Home Capital debacle at the end of the month, it was smooth sailing for the markets in April. Canadian credit continued to perform, led by higher-quality long bonds and solid ‘Triple B’ names. Bank NVCC has been the star performer narrowing 55 bps so far this year. We elected to take profits and exit our position and will wait patiently for a re-entry point.
We did not have any exposure to Home Capital, and when the OSC allegations became public, we reduced/exited positions in regional lenders. Exposure to the outperforming sectors coupled with carry contributed to a strong performance of 1.03% in April.
Home Capital continues to dominate our radar screen. They have three outstanding bond issues trading in the low ‘nineties’ including one that matures on May 24th. It is clear that people are worried about how this story plays out.
Their business model is based on funding long-term assets (mortgages) with large short-term borrowing (high-interest savings accounts and GICs). The recent run on deposits is a “loss of confidence” story, and it is hard to imagine how management can stem the withdrawals. As veterans of the 2008 financial crisis, we know that fear is contagious and can lead to knock-on effects impacting other lenders including the banks. Since the problem isn’t to do with the quality of the mortgages themselves, but rather a loss of confidence in Home Capital, it is likely that this mess is an isolated one.
Outside of the financial space, we expect the demand for credit to remain robust. Until there is some resolution to the Home Capital saga, we will maintain a medium risk posture.
In what seems to be the normal pattern, US first-quarter data was uninspiring. Furthermore, it is becoming apparent that President Trump’s ambitious agenda will likely unfold at a much slower pace than originally anticipated, and will likely be less aggressive than the platform he campaigned upon. As a result, US yields drifted roughly ten bps lower during the month. Even though Canadian GDP was quite good, many believe the results are transitory which took Canadian rates lower by the same amount.
The Federal Reserve has signaled that the US economy is expanding as expected so remains poised to lift rates as early as June. Meanwhile, the Bank of Canada will almost certainly leave rates unchanged and provide a cautious outlook when they meet on May 24th.
The Algonquin Team