SocGen’s Andrew Lapthorne is going to make you a believer in the idea that spreads are disconnected from measures of corporate leverage if he has to kick down your damn door and scream it at you.

Regular readers have probably some semblance of familiarity with Andrew’s work and he talks (a lot) about leverage and the risks inherent in lack of balance sheet discipline.

Perhaps our favorite Lapthorne quote is his straightforward assessment of companies who issue debt at artificially suppressed borrowing costs in order to implement bottom-line-inflating buybacks at nosebleed equity valuations:

As we have long pointed out, the reason for [the] increase in debt is largely down to financial engineering – aka share buybacks. Borrowing money to buy back your elevated shares is clearly nonsense.

Right. Or actually, it’s “clearly nonsense” if you care about doing the responsible thing, but it’s “clearly” awesome if what you care about your equity-linked compensation and adopt a myopic approach to management. Look on the bright side.

The other thing about buybacks is they act as real-life plunge protection for the broader market in a scenario where systematic strats are forced to de-risk and long investors are failing to BTFD. Just ask Goldman, whose buyback desk had its most active two weeks in history during the February rout. Recall this from a February 24 note:

Tax reform and the recent market correction will fuel 23% growth in buybacks to $650 billion. The Goldman Sachs Corporate Trading Desk recently completed the two most active weeks in its history and the desk’s executions have increased by almost 80% YTD vs. 2017.

Here’s a fun chart:

Goldman sees something on the order of $650 billion in buybacks in 2018 but JPMorgan is even more optimistic, calling for more than $800 billion in repurchases.

So you might remember that back in August, Lapthorne went looking for proof of whether buybacks in fact boost the shares of the companies who institute them and his conclusion was that there’s not much evidence for that contention. Recall this: