Go to Gazelle.ai

Greetings from the Chief Economist and Chief AI Officer of Gazelle.ai

Gazelle has recently received important and timely inquiries regarding whether there may be changes in the interpretation of our GScores during the current Covid-19 outbreak. For example, what does a 5 for the Hospitality and Services industry now describe? An important strength of our GScore, that was built in from the beginning, is its ability to distinguish higher performing firms in an industry regardless of the current market conditions, i.e. in both Boom times and Recessions. This is because the GScore is a relative measure distinguishing firms in in an industry. One implication of this is that 5’s in one industry are not the same as 5’s in another industry, or in a previous release. Still, the ratio of 5’s in an industry compared to the total number of firms in that industry does allow one to compare industries' performances; i.e. industries with a higher share of 5’s are doing better. 

The more important implication of this relativeness is that even if a slowdown impacts each industry uniquely, this effect is arguably more common within an industry, which is where the GScore is making its relative comparison. So 5's in a recession are still the best bet for finding an expanding firm even though a 5 now is not the same as a 5 two months ago. Similarly, in the reverse, firms with a GScore of 1 prior to Covid-9 are now at more risk. This is similar to using a curved grading scale at University. For example, an A in one class to the next may not reflect the same performance/skills/effort, but those A’s both still identify the top students across classes. 

Thus, the interpretation of those Hospitality sector 5’s above is that they are the strongest firms within that industry at this time. And even as more recent data for the industry comes in, e.g. decreased firms’ sales and/or supply chain disruptions, all else being equal the GScore scale will mostly just slide down reflecting all firms’ lower performances, excepting the firms that somehow insulated their sales and supply chains, leaving the top performers or 5’s mostly unchanged. As newer data arrives for the Covid-19 period, all else being equal, this will reshuffle some of the top firms, moving upward those with more insulated positions or in industries seeing higher demand (e.g. PPE firms), while everybody else shifts downward in mostly the same relative position, and with mostly the same GScores.

That being said, we are nevertheless taking a number of R&D steps to try and increase the sensitivity of our AI to short-run (weekly to monthly) transitory events such as Covid-19. This will allow us to better identify the extent to which firms may be insulated as described above. These improvements involve sourcing and incorporating more recent and higher frequency data. In addition to other new data included in the next release, weekly predictors we are now incorporating include: initial unemployment claims, and immediately previous VC funding events, acquisitions, and expansion events for the firm in question. Other high frequency data incorporated include: monthly electricity usage by state, monthly housing starts by state, changes in monthly employment by state and industry, changes in monthly exports and imports by state and industry, monthly national industrial production by industry, and average employment changes in up to the top 30 upstream suppliers and downstream buyer industries. 

These improvements allow us to better predict how firms may be differentially impacted by short-run market events, and will aid our clients in better identifying remaining opportunities, as well as helping development agencies identify and mitigate the negative effects on firms least insulated from Covid-19.

Did this answer your question?