During this recent webinar (https://refini.tv/37DJGNI) for business partner BattleFin, we examined some key performance indicators (KPIs) with an upcoming earnings season in mind. In my contribution to the event, I used BP (BP.L), formerly British Petroleum, as a running example.
One reason for choosing BP is that there’s been a very broad range of performance returns so far this year among oil majors, as seen in Exhibit 1. That suggests that any insights into future earnings that can be gleaned from data may be especially valuable in this industry segment.
Exhibit 1: Year-to-date Oil Majors Stock Price Performance
To overstate the obvious, a factor that always matters during earnings season is the consensus earnings estimate. The I/B/E/S Mean EPS estimate is the gold standard for benchmarking market expectations and determining if a company beat, met, or missed expectations when it does report.
Exhibit 2: I/B/E/S Mean and Median Estimates, StarMine SmartEstimate® and Predicted Surprise %
Source: Eikon by Refinitiv, I/B/E/S estimates and StarMine SmartEstimate® from Refinitiv
Included alongside the I/B/E/S estimates in the Eikon Detailed Estimates view shown in Exhibit 2 is the StarMine SmartEstimate from Refinitiv. StarMine is our predictive analytics and quant models brand, acquired in 2008. The SmartEstimate aims to provide a more accurate estimate by placing more weight on the most recent estimates and more accurate analysts. That formulation is shown in Exhibit 3.
Not typically blurred, the contributing firms and analyst names are displayed alongside their one- to five-star ratings, with 5-star rated analysts representing those whose estimates have been more accurate, early, and different from consensus in the past.
Two estimates are highlighted in red as being a “Bold Estimate.” Those are estimates from 5-star rated analysts that are significantly different (in this case lower) than the consensus estimate.
Exhibit 3: StarMine SmartEstimate Details and Estimate Weightings
Source: StarMine from Refinitiv
SmartEstimates are especially informative when they differ significantly from consensus (purple line in Exhibit 4). The percentage difference between the two is named the Predicted Surprise (PS%). When the SmartEstimate (shown in orange) is different by 2% or more, we find that companies report earnings surprises in the direction of the SmartEstimate approximately 70% of the time. Recall from Exhibit 2 that BP’s PS% is -31.01%.
Never have we observed such large PS% numbers as during recent months. We attribute this to economic displacements, both positive and negative, during the Covid-19 pandemic, withdrawal of guidance among many companies and, in the case of BP, an oil price war and demand shock. To help anticipate the direction of future revisions and earnings surprises, the SmartEstimate may now be more valuable than ever.
Exhibit 4: SmartEstimate, I/B/E/S Mean and Stock Price for BP, Fiscal Year 2020 Estimates
Source: StarMine from Refinitiv
The SmartEstimate is a valuable KPI for earnings season wherever you’re looking for a more accurate estimate. Its accompanying Predicted Surprise% is incorporated in the StarMine Analyst Revisions Model (ARM), shown here for BP in Exhibit 5. Estimate revisions and the PS% are incorporated across the income statement and for multiple periods: this quarter, this year and next year.
This model measures changes in analyst sentiment and is predictive of the direction of future revisions, which in turn, is correlated with future price changes. Over the trailing 12-month period through June, this model has delivered top/bottom decile spreads of 9.3% across a universe of the top 2,000 stocks by market cap in developed Europe and 13.4% among the top 600 UK-listed stocks.
Exhibit 5: BP’s StarMine Analyst Revisions Model Details
Source: StarMine from Refinitiv
Industry-specific KPIs provide additional information. We generate hundreds, including with them a SmartEstimate and Predicted Surprise%. Exhibit 6 shows those for BP’s Oil Production Per Day estimates.
Exhibit 6: 2020 Oil Production Per Day Estimate
Source: StarMine from Refinitiv
Exhibit 7 shows BP’s SmartEstimate for oil production moved sharply below the consensus on April 30.
Exhibit 7: History Chart of Oil Production Per Day Estimates and Stock Price
Source: StarMine from Refinitiv
Refer back to the EPS estimates history chart shown in Exhibit 4. In the version shown here in Exhibit 8, note the large drop in the SmartEstimate on April 3, followed by a series of further downgrades in both it and consensus. Keep that date in mind.
Exhibit 8: SmartEstimate, I/B/E/S Mean and Stock Price for BP, Fiscal Year 2020 Estimates
Source: StarMine from Refinitiv
In Exhibit 9, we join data from third-party alternative data provider, Minneapolis based LinkUp, whose job market data is now on our business partner’s BattleFin Ensemble Platform and soon in our own quant platform, QAC (QA Direct in the Cloud). This is shown alongside Refinitiv’s own I/B/E/S and SmartEstimates EPS data. The purple line shows the consensus, or I/B/E/S Mean Estimate. The orange line represents the StarMine SmartEstimate. Observe the SmartEstimate leading the consensus lower.
Among the factors that matter going into and throughout earnings season, alternative data lends an increasingly valuable source of intelligence.
In this example, LinkUp showed BP job postings being pulled down prior to the series of estimate cuts. The number of active job postings provided by LinkUp is shown in yellow. That number hit a high of roughly 800 in early February before beginning to decline in early March and being aggressively taken down later that month. This nearly total hiring freeze was a harbinger of the estimate cuts we observed in early April.
Without this aggregated job postings data, it would have been exceedingly difficult and unlikely that a fundamental analyst would have been able to notice them being taken down.
Exhibit 9: BP Number of Active Job Postings and EPS Estimates, Jan. 1 to June 25, 2020
Sources: StarMine and I/B/E/S data from Refinitiv, LinkUp data from QA Direct
Conclusion
A combination of I/B/E/S EPS, and industry-specific KPI estimates, the StarMine SmartEstimates and Predicted Surprise analytics, and sources of alternative data should provide investment professionals with insights that better equip them for future earnings seasons.
During our recent BattleFin webinar, we provided a step-by-step demo of how to access and extract LinkUp data from within our QA Cloud quant platform. You may view that here: https://refini.tv/37DJGNI
