by Tim Gaumer.
The consumer technology industry is facing headwinds. Consumer demand appears to be weakening alongside an economic slowdown. Channel inventories are on the rise. Covid-19 related lockdowns are creating supply chain problems within and from China. Higher semiconductor and component prices are increasing margin pressures and a stronger U.S. dollar isn’t helping.
In this update, we’ll demonstrate how these developments have changed the sell-side outlook for two companies: Tokyo Stock Exchange listed Sharp Corp (6753.T) and Hong Kong listed Lenovo Group Ltd (0992.HK) and how the StarMine Analyst Revisions Model (ARM) from Refinitiv provided a rapid signal to users of significant changes in analyst sentiment, as represented by estimate revisions and/or recommendation changes.
ARM is a stock-ranking model designed to predict the direction of future analyst revisions. This is valuable to investors since revisions tend to have a positive correlation with the direction of future price changes.
Sharp Corp. (6753.T)
Exhibit 1 shows the ARM score for Sharp on September 26th. Its score of 10 is a regional-relative percentile rank. So, its listing is ranked in the 10th percentile (bottom decile) relative to all other companies in the Developed Asia Pacific region. Like all StarMine quantitative model scores, stocks are ranked between 1 (worst) and 100 (best). High scores are a more bullish signal.
Exhibit 1: Sharp Corp’s ARM score on September 26, 2022
Unlike a basic revisions model, ARM looks across the Income Statement at changes not only in EPS estimates, but also EBITDA (or, for Japanese companies, Operating Profit) and Revenue. It also looks at changes in analyst recommendations. It does this over multiple timeframes and for estimates over multiple fiscal periods: typically for the current quarter, full-year, and next-year. These details can be seen in Exhibit 3.
All StarMine model detail pages are accompanied by a history chart that shows both changes in the stock price and in the model scores. You can see in Exhibit 2 that Sharp experienced a negative and persistent series of analyst revisions beginning in early summer, with the first big drop in its ARM score occurring in May, 2022.
Exhibit 2: Sharp’s ARM 3-Year History Chart
Looking at the model components view in Exhibit 3, we see major downgrades in Operating Profit and EPS estimates for the current quarter, full fiscal year, and next year. The StarMine SmartEstimate is lower still, resulting in large Predicted Surprise percentages. Each of the ARM input components gets their own percentile rank scores and these large estimate cuts places both EPS and Operating Profit measures into the bottom decile relative to all other companies in Sharp’s region.
Negative model inputs are highlighted in red, neutral in grey, and positive inputs, were there any, in green.
ARM incorporates the StarMine Predicted Surprise analytic as a reinforcing signal. Predicted Surprise is the percentage difference between the StarMine SmartEstimate and consensus (the I/B/E/S Mean estimate). Whereas the consensus estimate places equal weight on every analyst estimate, the SmartEstimate reweights the average to place more weight on the more recent estimates and the more accurate analysts (see Exhibit 5).
In the time-series charts above, the SmartEstimate is shown as the gold line versus the Mean estimate in blue.
The Predicted Surprise% (PS%) is predictive of the direction of future changes in the consensus, as lagging analysts catch up to those more responsive analysts. Further, a PS% greater than 2% or less than -2% is deemed ‘significant’ as our research shows that it will accurately predict the direction of a subsequent earnings surprise 70% of the time.
Exhibit 3: ARM detailed components view for Sharp
Exhibit 4 is from the Workspace Detailed Estimates view. This shows what is driving the negative EPS score for Sharp’s full fiscal year. The Mean (consensus) estimate is 70.293 Yen, while the SmartEstimate is 62.311 Yen, resulting in a Predicted Surprise of -11.35%.
Within the last 30 days, three new estimates have been published, with an average reduction of nearly 16% below their prior estimates. This has driven the Mean estimate 5.19% lower. The stock price has followed, moving 12% lower.
Exhibit 4: Sharp’s FY Mar-23 full-year EPS estimates
Exhibit 5 is from the same Detailed Estimates page. In this view, you can see the details behind the SmartEstimate. Greater weight is assigned to the most accurate (4- and 5-star) analysts and the most recent estimates. Estimates older than 120 days are marked X-Age and excluded from the SmartEstimate.
We deem an estimate from a 5-star analyst that is significantly different than consensus to be a ‘Bold Estimate’ and highlight negative Bold Estimates in Red. A positive Bold Estimate would be highlighted in green.
Exhibit 5: FY-1 Detailed Estimates
Exhibit 6 zooms in on these current estimates. Within Workspace, the name of the contributing firm and the name of each analyst covering Sharp would be visible to the user. Here, it’s easier to see that one of the most accurate 5-star analysts has published an EPS estimate of 44.650 Yen. That’s 36.48% lower than consensus. It is somewhat unusual for analysts to stick their neck out to this extend, which is why it has been tagged a Bold Estimate.
Bold Estimates can be screened for across a larger universe of companies in the Workspace Screener App.
Exhibit 6: Expanded view of current estimates
Lenovo Group Ltd. (0992.HK)
Lenovo Group also shows a negative change in analyst sentiment, as highlighted by its StarMine ARM score. Exhibit 7 shows the ARM score for Lenovo on September 26th. Its Hong Kong listing is ranked in the 8th percentile relative to all other companies in the Developed Asia Pacific region. Lenovo hasn’t always had a low ARM score – it was in neutral territory as recently as last week.
Exhibit 7: Lenovo’s ARM score on September 26, 2022
ARM looks across the Income Statement at changes in Lenovo’s EPS, EBITDA, and Revenue estimates over multiple timeframes and fiscal periods. It also looks at changes in analyst recommendations. These details can be seen in Exhibit 8.
Negative model inputs are highlighted in red, neutral in grey, and positive inputs, were there any, in green. Analysts have been cutting their forecasts for revenue, EBITDA and EPS, especially for the full fiscal year (ending March, 2023) and for the following fiscal year.
Estimate cuts across the income statement have resulted in each of the revisions component scores to drop to the bottom quartile. These low scores are further driven by the negative Predicted Surprise% numbers that accompany downward revisions to the Mean estimates.
Adding to the overall ARM score of 8 has been negative changes in analysts’ recommendations, resulting in that component’s score of 19. The Mean (average) Recommendation is on a scale of 1-5, where 1 represents a Strong Buy recommendation and 5 a Strong Sell. More predictive of future stock price performance is the change in recommendations, rather than the level. The 0.8 increase in the Mean Recommendation indicates that the overall average recommendation has moved closer to Hold than in the opposite direction.
Exhibit 8: Detailed ARM Component view for Lenovo
Several factors have recently conspired and thrown up stumbling blocks for at least a couple of Asia-based technology hardware vendors. StarMine ARM is an effective and time-savings predictive stock-ranking tool for helping users quickly discover fundamental changes that have caused the analyst community to significantly change their outlook on a company.
For more on this, please see our September 20th Product Insight on FedEx or request a copy of our white paper.
Refinitiv Workspace is a complete solution for research and analytics. It places the most comprehensive market information, news, analytics and trading tools available into a desktop.
Refinitiv I/B/E/S Estimates are a market leader, boasting 200+ metrics and indicators across 15 industries. Find more information on our estimates data.
Get unique value-add analytics and predictive financial modeling, dedicated to making investment research smarter with Refinitiv StarMine data.