JPMorgan: You can profit 70% of the time you buy the dip in corporate debt

Investors who buy high-quality U.S. corporate bonds when spreads widen have made profits over the next three months about 70% of the time, strategists led by Eric Benstein and Nathaniel Rosenbaum wrote Thursday.

“Historically, it seems relatively clear that most dips in HG are meant to be bought in the near term,” the strategists wrote.

Spreads on high-grade U.S. corporate bonds rose in August but have since partially recovered. After averaging about 92 basis points, or 0.92 percentage points, in the first seven months of the year, they widened to 111 basis points on Aug. 5. They have since stabilized at 100 basis points as of Wednesday, according to Bloomberg data.

Strategists looked at sell-offs in the JPMorgan U.S. Liquid Index, or JULI, an index of investment-grade companies. They analyzed times when spreads were at their widest in three months and remained at their widest point into the following month. They also looked at periods when the maximum spread was about 15 basis points wider than the narrowest spread in the previous three months, to ensure that the moves were at least moderate sell-offs.

By that definition, there have been 37 sell-offs since 2000. And if someone bought at the widest point, when the pattern worked, the subsequent narrow level was on average about 46 basis points narrower over the next three months, the strategists wrote.

But there have been instances where these tactics have not worked. On 11 occasions, a larger selloff occurred three months later, with the market widening by at least five basis points. In May 2022, spreads widened to 173 basis points, then narrowed, then suffered another selloff two months later, to 180 basis points, as the market mispriced the Fed’s expectations of a rate hike.

The analysis is primarily useful for giving a sense of history, not as a trading strategy, because investors don’t know in the middle of a sell-off when the market is at its widest point, the strategists wrote.