April 11, 2020

Coronavirus, Correlations, and Chicken Consumption

Writing more is one of the habits I’d really like to work on during this quarantine season, extending the list of other noble habits I’ve developed: consuming a Haagen-Dazs Vanilla Milk Almond Ice Cream Bar every other day; binge-watching Conan; managing to eat, sleep, work, read, even workout all in my bedroom (that’s not meant to be impressive but highlight how slothlike I’ve become). I’m hoping to sit myself down and write every weekend. The content of this post will deviate tremendously from posts I’ve put out in the past, and it’s my hope to keep the content varied. You’ll see anything from personal journaling to data analysis to creative writing.

Coronavirus has understandably been on everyone’s mind, and there have recently been talks of the relationship (and even trade-offs) between public health and the economy so I wanted to visualize it. Plotting the number of confirmed cases over time from Johns Hopkins’ COVID-19 dataset, you get something like this.

Not so fun but also not that telling – all you see is a hockey stick graph but graphs like these are not particularly informative. Let’s come back to this later and plot the S&P 500 next since its a solid representation of the overall market and a leading economic indicator of how the US economy is doing.

So it looks like the US economy isn’t doing too hot but you didn’t need me to tell you that. (It’s also been climbing recently, rallying 12% this past week and up ~25% since its March low, so is this resurgence fools’ gold or the real deal?) Next, let’s plot the number of COVID-19 confirmed cases against the S&P 500.

Pretty neat – seems like the market has been trending down as COVID-19 cases rise but, again, not very telling.

Here’s where I think it gets incredibly fascinating and more telling though. If you, instead, calculate the weekly percent change in COVID-19 confirmed cases and the S&P 500 and then plot both, you get a stunningly negatively-correlated figure.

You can’t make this up – this is about as good as it gets in the wild. Like between this signal is so bad it makes AT&T’s coverage in a Cold War underground bunker look decent and this signal is so good that not only am I investing my life savings but my parents’ too, this signal is like WiFi in my bedroom: a solid 7. Rarely will you be able to plot two not-immediately-related concepts and see such a clear visual relationship. This is nothing like the smoking-and-lung-cancer or fast-food-and-obesity kind of correlation, where an obvious, intrinsic relationship exists between the two concepts. Think about this for a moment – one line shows us the weekly change in how many people in the US are getting sick with the coronavirus and the other line shows us the weekly change in how a specific stock index is doing. Let’s flip the S&P 500 weekly change curve upside down just to see how closely in-sync these two time-series are.

There’s a lot of market pessimism right now with a considerable number of folks expecting another plunge. However, note how – more recently – the weekly change in the S&P 500 has climbed while the weekly change in US COVID-19 cases has gone down. Does the market know something that we don’t? Is there more pessimism than there should be and is the market rightfully optimistic?

The simple answer is this pessimism rightfully exists. For instance, maybe dramatic but Steve LeVine suggests a depression is looming in his Medium post. Take a look at the federal funds rate, the 10-year treasury bond rate, and the S&P 500 all plotted over time since 1965, with recessionary periods since then shaded a darker gray.

We just want to check our intuition here – the S&P 500 is a leading indicator of how the economy is doing and, yeah, looks like every recession involves a dip in the S&P 500. Also, the 10-year treasury bond rate and the fed funds rate move more or less together as expected.

The 10-year treasury bond rate is how much interest you would receive if you purchased a 10-year treasury note – a loan to the US government that matures in a decade. It’s the most popular debt instrument in the world, and it’s often considered one of the best indicators since the more optimistic investors are in the economic future of the US in the next decade, the higher the yield (or interest rate). There’s been a steady slide in the last several decades, and it’s been at a record low as of late. Lastly, the federal funds rate, the interest rate at which banks lend from reserve balances to one another overnight – is often lowered by the Federal Reserve to stimulate the economy. I’m sure you’ve heard about how it’s plunged to a near-zero rate recently. Combined with how the Federal Reserve has also done some quantitative easing, the US has brought out its big guns to this economic war. I’ve recently heard this situation described as a fleet of firetrucks – the firetrucks tell you significant effort is being made to fight the fire, but this also means the fire is significantly large.

There’s no telling where the economy will head, and my speculation is probably worth a taco truck taco at most – something I desperately wouldn’t mind having and eating right now – but I found it fascinating to think about and visualize how this moment is both similar to periods in the past and something we’ve never seen before. One thing is for certain though: public health and personal health are and should be of utmost priority.

And just to keep things light and show that misleading, spurious correlations do exist, here’s the annual per capita poultry consumption in the US against the S&P 500.