One of the constant pressure during my career in financing has been the drive to save costs. This inevitably leads to a man in my company who asks me to downgrade or to report from Bloomberg to FactSet for specialized data providers. The Bonentellers do not realize that if you work in investment research or achieve assets, these data subscriptions can make a difference between good and poor performance and therefore between getting more customers (and higher income) and losing customers (sorry for the mini -rant).
Braiden Coleman, Travis Dyer and Mark Lang investigated more than 200,000 analyst reports from stock agents and tried to find out what effect the data sources that these analysts use in their reports have. The graph below is a mix of three different regressions that I have contracted in one graph. Moreover, the researchers unfortunately always place each variable in deciles, so that the figures cannot be translated directly into an X% rise or decrease in accuracy. However, it enables us to compare the effect size of data subscriptions with other factors.
Effect of data subscriptions on the accuracy of the analysts forecast
Source: Coleman et al. (2025)
Let’s look at the four bars to the left of the graph, which are derived from their basic results. We can see that the number of data subscriptions influences the accuracy of analysts more prediction than analyst experience. A young analyst with many different data sources will be better than the 20% most experienced analysts.
On the other hand, the graph also shows that analysts at larger brokerage companies are usually less accurate than analysts at smaller houses – something that I attribute to the juniorization In large companies, where experienced but expensive analysts are replaced for years by younger, cheaper.
Let’s go to the right half of the graph, where I have brought two separate regressions together in one graph. We can see that the five major data providers (Bloomberg, S&P Capital IQ, Factset, LSEG, Morningstar) offer some boost for the accuracy of the analysts. But smaller data providers with more specialized data have a 60% greater impact on the accuracy of analysts forecast. In the meantime, publicly available data sources offer only limited benefits.
But unfortunately these smaller, more specialized data providers are often canceled before the large one. That means that cost savings take place where analysts tend to get the biggest value for money when they do their work.
Even worse, the study also finds that analysts who use the same sources as other analysts tend to make similar predictions. Since the cost savings in our industry are leaving more and more of us with one of the five large data providers and nothing else, this means that we lose our ability to distinguish ourselves on the market. And that means that there is less independent thinking and more group thinking in markets. Plus, if you are an investor who has to pay for research into the sales side (thank you, MiFID), why would you pay for research that is not differentiated? You can easily stop paying for the research and using consensus data, which are available with the most important data providers such as Bloomberg or LSEG workspace.
In short, reducing the costs by removing the access of analysts and investors to data is self -destructive. It only speeds up the downward spiral, which can lead to lower income and more cost -saving in the long term. Like I said before You cannot give your way to profitability.
#Data #important