We designed Katana to help investment professionals separate the signals from the noise as prices change over time. Since we released our first prototype almost 2 years ago, Katana has helped portfolio managers find opportunities they would otherwise miss. In a mild volatility environment Katana detects the rare anomalies that demand attention; those changes in relative spread that are statistically significant. We have seen before how Katana identifies local events like the Turkey crisis in 2018 and the riots in Chile in October 2019. With the outbreak of COVID-19 triggering extreme turmoil in the market, Katana is giving us new insights into how relative value evolves in a high-volatility environment and an opportunity to train and test our algorithms in a new market regime.
Let’s start with the big picture. While in a normal market Katana finds around 100 relevant dislocations per day, that number has jumped to more than 400 over the last week. This is in line with our intuition that the number of dislocations in relative value across the bond universe scales with the average dislocation. In other words, the more the market as a whole moves, the higher the disparity in the speed at which spreads move. With Katana we can accurately measure what is moving faster and what is lagging behind.
Some of the largest relative movements in EM Credit at the close of Tuesday March 18 have been among Financials, Healthcare, and Energy.
In particular Katana highlights:
- Issuers in the oil industry in Mexico and Indonesia suffered the largest relative losses.
- Brazilian issuers in transportation underperformed.
- There has been a large sell off of Ukraine against Russia.
Overall, the largest number of dislocations are in Mexico, with 158 trade ideas, followed by Brazil with 100. In both cases, the dislocations are primarily happening between different issuers within each country; 153 in Mexico and 76 in Brazil. In Mexico, all but 9 trade ideas show Pemex getting cheaper across the curve. The belly of the curve has got cheaper against the sovereign as well as the rest of the corporate sector generally. The majority of ideas are in the 6-10 years maturities (119), but there are no dislocations of a similar scale within the Pemex curve itself. For example, the chart below shows Pemex 27 against Mexico 26, where the spread difference has spiked to 582 bps, almost 7 standard deviations higher than the average over the last year of 264 bps.
In Brazil the biggest dislocation has been in Azul Investments (AZULBZ), which services the airline industry, and accounts for 31 of the 76 ideas, with the 2024 bonds widening to over 2,100 bps. It is followed by another company in the logistics sector, Rumo (RAILBZ), which operates railways, and looks cheap against 11 corporates and sovereign bonds by over 5 standard deviations.
Unlike those in the 2 largest markets in Latin America, where the anomalies are primarily internal, Indonesian corporate bonds have mostly gotten cheaper against those in other countries. Only 10 out of 54 trade ideas are between Indonesian issuers. Medco Strait Services (MEDCIJ), in the Oil Exploration and Production sector, has sold off significantly more than 6 other Indonesian corporates and 18 issuers from other countries that it has been correlated with in the past.
Both sovereign and corporate bonds in Ukraine have been widening faster than Russian ones. At the sovereign level, Ukrain 2028 bonds have widened to 1,76 bps, 797 bps over the Russia 2027 and almost 6 standard deviations from the mean.
A challenging trading environment increases the need for sharper tools. Request your demo of Katana below and start a free trial. We already cover EUR IG as well as EM Credit and will soon add more asset classes.