London (HedgeNordic) – Visible Alpha began in 2011 as a project backed by a consortium of banks who still support the firm. The company was carved out in 2015 and has now grown to over 800 employees globally. Demand for equity research actually increased after MiFID II unbundling required explicit payments for it. “Investors look at sell-side analysis and forecasts because no other source is as broad and deep for proving thoughtful and detailed analysis,” says founder and Chief Research and Innovation Officer, Scott Rosen. But today’s investors are increasingly demanding to see the rationale underlying the research.

“Investors look at sell-side analysis and forecasts because no other source is as broad and deep for proving thoughtful and detailed analysis.”

When Rosen was a sell-side equity analyst in the 1990s, he found, “analysts had a huge information advantage in terms of access to information and could simply regurgitate it. Now, the environment has changed to a level playing field and investors are looking for insights to understand what is behind analyst views. But the difficulty of comparing analysts on very detailed issues created a granularity disconnect where it was hard to get consistent information about assumptions behind forecasts.”

Visible Alpha’s “deep consensus data” addresses this buy-side problem by demystifying models that were once perceived as “black boxes.” The firm’s platform is comprised of a library of historical and forecast data, including analysts’ own estimates of geographic, segmental and product breakdowns (which may not be publicly disclosed by companies). “Visible Alpha builds consensus estimates from the bottom up, normalizing the data inputs, rather than just extracting and aggregating headline conclusions,” says Erin Gifford, Marketing Director. The average number of line items feeding into a forecast is 161, but there could be as many as 1,000 for some companies. Visible Alpha is agnostic on accounting systems, but makes the data comparable — US GAAP and IAS are most often used in the raw estimates. “We aggregate what would otherwise be very disparate information,” says Rosen.

“Visible Alpha builds consensus estimates from the bottom up, normalizing the data inputs, rather than just extracting and aggregating headline conclusions.”

The most intense users of the data tend to be fundamental equity long/short analysts feeding ideas into concentrated books at hedge funds. There are also a growing number of long only and quant clients. Some users are credit investors, but Visible Alpha does not currently capture models from fixed income analysts. “The market for fixed income research is much smaller and we would struggle to find enough names that have 3 analysts with enough granularity and similarity,” points out Rosen.

Use Cases

The word “consensus” is misused by the media to imply an agreement distilled down to one definitive figure, when there are in fact multiple ways to define and measure the concept of “consensus” – and there can be differences between Bloomberg, Refinitiv and Visible Alpha consensus metrics. “It is a complete misnomer to imply agreement. If there was agreement, we would not need the market. In fact, there are a variety of forecasts and disagreements,” explains Rosen.

“The consensus is just the mean summary statistic and arithmetic point of various numbers and this can be the least interesting part. Knowing the range of views can be more useful.”

“The consensus is just the mean summary statistic and arithmetic point of various numbers and this can be the least interesting part. Knowing the range of views can be more useful. Sometimes there is very tight agreement and sometimes a very broad spectrum of views.” Clients can see the mean, mode, range and standard deviation of the estimates. Outliers are shown and could be excluded to create a different measure of consensus.

Data can be used for forecasting and also benchmarking opinions. “Consensus estimates can be a proxy for forecasting for those who do not have time to dig deep into the dynamics of the company. Alternatively, consensus estimates can be used to represent what the market believes and what is already reflected in current pricing. That can provide scope for expressing contrarian views on pricing, revenues, or other variables,” says Rosen.

Investors might employ the data to find the fastest-growing companies in a sector that is doing well, such as AI. Or they could look for companies breaking out within an industry that is not doing well, such as traditional retail. The data could also feed into pairs trades for long/short or market-neutral strategies.

Visible Alpha produced an ad hoc report, “Shopping for Alpha,” illustrating a possible investment application of picking winners and losers within sectors. It showed how two carefully selected metrics – comparable sales and operating margin growth – could have been used to predict relative share price performance within two highly leveraged industries: retail and restaurants. The metrics were chosen to enable apples-to-apples comparisons over time and between companies.

Accuracy and Valuation

Visible Alpha’s standard output equally weights the latest analyst models and does not attempt to rank individual analysts on accuracy of forecasts, though a client could, to some degree, create their own customized consensus subject to brokers entitling the client to their specific data.

“The primary role of analysts is understanding the mechanics of a business and how a company weathers change in a competitive landscape…”

A reversal from bullish to bearish analyst sentiment (or vice versa) can be of most interest to institutional investors. It is important, however, to distinguish between the stronger and weaker skillsets of the typical sell-side analyst. “The primary role of analysts is understanding the mechanics of a business and how a company weathers change in a competitive landscape, over long-term forecasts of 1- 2, or even 10-20 years. They are less good at working out what the market is willing to pay, because valuations depend on macro-market psychology. Big US brokers may give risk assessments but do not give so many short-term trading recommendations,” says Rosen.

There can also be a perennially optimistic bullish bias in estimates, though accuracy checks occur every quarter when results are released.

Updates and Corporate Events

Analyst models need to be updated after key events such as earnings releases to keep their place in Visible Alpha consensus estimates. Earnings are the biggest reason for updating estimates, but analysts do update models in the interim throughout each quarter. A mix of machine learning and teams of sector specialists at Visible Alpha manage the updating process.

Corporate events can be another reason for updates. The Visible Alpha broker universe does not include specialist event-driven houses, who broke to merger arbitrage and other event funds and do not cover companies on an ongoing basis. However, “regular sell-side research around merger events may add another version of the future with proforma estimates if a takeover is completed,” says Rosen.

Company and Broker Coverage

At least three analysts need to be covering a stock for Visible Alpha to include it. Coverage is currently around 6,500 companies globally, which are mainly large-and mid-cap names. Coverage of small and micro caps is being expanded as Visible Alpha develops relationships with more brokers.

Most of the 180-strong Visible Alpha broker universe is listed on their website, and has been growing in Asia after the firm opened an APAC office in Hong Kong in 2019. It does not include specialist short-selling research houses, such as Hindenburg, because their published research tends to be more opportunistic and ephemeral in nature. “We like to see longevity of coverage, with some analysts having followed a stock for 20 or more years,” says Rosen.

“We like to see longevity of coverage, with some analysts having followed a stock for 20 or more years.”

The list of brokers does include some independent research providers (IRPs), including some bank and broker units that have been reclassified as such post MiFID II.

A multitude of public companies citing Visible Alpha data, as well as buy-side users encouraging the sell side to contribute, are helpful in increasing contribution and coverage. One incentive is that sell-side contributors get free access to consensus data, and brokers can also receive some remuneration for their efforts.

Visible Alpha’s business model is primarily Software As A Service (SAAS) subscriptions based on a range of metrics, including the number of users. “All clients essentially access the same data, with the exception of permissioned access to individual broker content. The client base also includes some smaller firms and even others that are pre-launch,” says Gifford.

Most clients use the web app, though there are also Excel add-ins and an API cloud model. Visible Alpha is not currently integrated with Chat GPT.