Working harder or working smarter; what helps analysts?

Nobody likes to be controlled and monitored by outsiders or their bosses. Especially not if you claim that in your job you work long hours. So, you can imagine how I felt when I learned that Azi Ben-Rephael and his colleagues monitored the average work length of sell-side analysts like me and my colleagues.

In their study, they exploited the fact that my fellow sell-side analysts and I are on Bloomberg virtually all day while we are working. So, they went to Bloomberg and asked for the records of the Bloomberg status of a bunch of sell-side analysts in New York City. They ended up with data for 336 analysts from 42 firms covering the time from the third quarter of 2017 to the first quarter of 2021.

So here are the bare bones facts:

The average workday of a sell-side analyst is some 10 hours long, but during the initial lockdowns in the pandemic in spring 2020, it increased to more than 11 hours as analysts no longer had to commute to work.

So far so boring. If we look beyond the facts, it gets interesting. People can spend a lot of time in the office but that may not necessarily mean they produce more reports or better analyses. It turns out that people who work more produce more output. A one-hour increase in the average work length increases the number of quarterly earnings forecasts made in notes by 5%. Analysts who work longer hours also publish their forecasts in a timelier manner after a company presents results. Updates to forecasts are published about 6% faster when a sell-side analyst works one hour more per day. And to top it all off, the earnings forecasts of analysts who work longer hours tend to be more accurate than the forecasts of analysts who work shorter hours.

Ok, so ‘working harder’ pays off for sell-side analysts. But what about analysts who log on to Bloomberg but then go to meet companies and other people, so their terminal remains idle for large parts of the day? They might actually ‘work smarter’ because they rely less on publicly available information and more on the information gathered from informal chats with company management, investors, and industry insiders. And that information may be more valuable than all the numbers of a Bloomberg terminal.

The researchers measured the percentage of time away from their Bloomberg terminal to estimate the time analysts spend in meetings gathering soft information and meeting people. This kind of soft information could be useful if analysts learn things from these meetings that help them make better forecasts.

On average, analysts spent some 28% of their time away from their desks but there were large differences between analysts. The range covering 80% of analysts was between 3.3% of their day away from the desk to 66% of their time away.

Being away from the desk significantly reduces the output of an analyst. Analysts with above median time away from the desk produce some 12% fewer earnings forecasts than analysts who spend more time on their desk than the average analyst. But it seems that time away from the desk means better information from companies. The forecast accuracy of analysts spending more time away from the desk during the day improves to a similar degree as the forecast accuracy of analysts who work longer hours. This indicates that both ‘working harder’ and ‘working smarter’ pays off for analysts.

Until a pandemic hits and you cannot meet people in person anymore…

The pandemic brought with it a unique test case for the impact of remote work. Naturally, as people worked from home they saved on commuting time and as a result, the average hours worked per day increased significantly (from 10 hours to more than 11 hours). But only the analysts who already worked long hours benefitted from that additional time at work. Their output increased and the accuracy of their forecasts increased roughly in line with the general improvements seen by people working longer hours.

Meanwhile, the analysts who typically spend a lot of time away from their desks talking to management and investors couldn’t do so anymore. On the plus side that gave them more time to publish notes and the raw number of forecasts made increased significantly. But the quality of their work suffered. A lot. The accuracy of the forecasts of analysts who normally spend a lot of time away from their desks dropped significantly.

What do we learn from these results? We learn that analysts who work harder produce better material. But analysts who rely more on soft information and meetings with other people can only gain an edge if they are physically in the office and meet people. Once they start working from home the quality of their work drops because they have less access to the main driver of their performance – the information gathered in meetings and discussions with other people.