Refinitiv
3 Times Square, New York, NY 10036
3 Times Square, New York, NY 10036
An update on the performance of StarMine SmartEstimate from Refinitiv and Predicted Surprise
To access other white papers, select from below and continue:
Performance of StarMine SmartEstimate from Refinitiv and Predicted Surprise for Developed Asia Ex-Japan
Synopsis:
StarMine’s proprietary SmartEstimate earnings prediction service was designed to create a better earnings forecast than the consensus estimate by differentially weighting analyst estimates based on each analyst’s historical track record and how long ago an estimate was issued. Recently, we published an investigation of the StarMine® SmartEstimates® from Refinitiv and Predicted Surprise success rates globally and by region for two different periods: from 1/1/1998 until 11/30/2008 and from 12/1/2008 until 11/30/2017. The current study drills down from those global numbers into the performance for the four countries in Developed Asia ex-Japan.
Please note that the document is an approximation of the original document.
StarMine SmartEconomics White Paper
Synopsis:
StarMine SmartEconomics applies StarMine SmartEstimates® methodology to forecasts of macroeconomic data and FX rates to create a SmartEstimate of economic data that is more accurate than the simple consensus forecast. StarMine assesses the historical accuracy of each contributor at every point in time on every poll in which the contributor had a forecast.
The indicator-specific StarMine historical accuracy score for each forecaster then determines the weight that each forecast receives in the SmartEstimate. Backtests show that the SmartEstimate for Economics and FX correctly predicts the direction of macro surprises relative to the consensus forecast about 61% of the time when the SmartEstimate is significantly different from the consensus.
Please note that the document is an approximation of the original document.
StarMine Earnings Quality (EQ) model White Paper
Synopsis:
The Earnings Quality model is a percentile (1-100) ranking of stocks based on sustainability of earnings, with 100 representing the highest rank.
StarMine defines earnings quality as a measure of the degree to which past earnings are reliable and are likely to persist. High quality earnings accurately reflect a company's current and past operating performance, are indicative of future operating performance, and are reliable valuation measures for the company, regardless of the level of earnings. Companies with poor earnings quality are not necessarily engaging in earnings manipulation; in most cases, low earnings quality reflects a likelihood of deteriorating fundamentals relative to the past. Furthermore, earnings quality can be measurably high: companies with very persistent earnings have strengthening fundamentals and are likely to outperform their benchmarks in the future.
Please note that the document is an approximation of the original document.
StarMine Analyst Revisions model (ARM) White Paper
Synopsis:
The Analyst Revisions Model (ARM) is a percentile (1-100) ranking of stocks based on changes in analyst sentiment, with 100 representing the highest rank. ARM is highly predictive of relative price movement and is effective across stocks in each capitalization category, investment style, and market sector.
Please note that the document is an approximation of the original document.
StarMine Price Momentum model (Price Mo) White Paper
Synopsis:
The StarMine Price Momentum ("Price Mo") model is a percentile (1-100) ranking of stocks based on recent historical price performance.
Higher scores indicate stocks with the strongest price momentum. An overall score of 95 indicates that the security has better price momentum than roughly 95% of its peers. The overall score for a given stock is a blend of its scores on each component. The headline StarMine Price Mo score featured in StarMine Professional is the Regional Rank, in which a given stock is ranked against all others in its Region. The StarMine Price Mo Global Rank is also available.
Please note that the document is an approximation of the original document.
StarMine Intrinsic Valuation model (IV) White Paper
Synopsis:
The process of determining the intrinsic value of a company entails discounting an infinite stream of some measure of future profitability (e.g., dividends, residual earnings, cash flow) to determine its true value. A proper intrinsic value calculation takes into account both the company’s current financial situation and its prospects for growing or declining profitability in the future. Most analysts and value investors believe that intrinsic value provides the benchmark toward which the stock’s price should be heading. However, not everyone agrees on what that intrinsic value should be.
StarMine chose a dividend discount model because it allows us to leverage one of our primary strengths—forecasting earnings. We have found that analysts estimate financial measures with varying degrees of accuracy. In particular, the accuracy of analyst estimates decreases from revenue, to EBITDA, to EPS, and then to cash flow. The cash flow estimates have approximately twice the error of EPS estimates.
StarMine’s Intrinsic Valuation Model begins with the sell-side analysts’ EPS and Long Term Growth (LTG) estimates. Our research has found that analyst estimates reflect systematic biases and that by accounting for the predictable nature of these biases, we can improve the overall accuracy of the forecasts. We make adjustments to the analyst forecasts and construct 15 years of SmartGrowth earnings projections for each company.
Please note that the document is an approximation of the original document.
StarMine Relative Valuation model (RV) White Paper
Synopsis:
The Relative Valuation (RV) model is a percentile (1-100) ranking of stocks based on price and enterprise value multiples.
Higher scores indicate the best value stocks, and correspond to the stocks with the highest yields. A score of 95 on the model's P/E Component, for example, means that the security has an earnings yield higher than roughly 95% of its peers. RV is highly predictive of relative price movement and is effective across stocks in each capitalization category, investment style, and market sector. The Relative Valuation model features six prominent valuation measures that are relevant to most securities: EV/Sales, EV/EBITDA, P/E, Price/CFO, Price/Book, and Dividend Yield. For companies that do not pay dividends, the model also considers share buyback activity.
Please note that the document is an approximation of the original document.
StarMine Value-Momentum model (Val-Mo) White Paper
Synopsis:
The StarMine Value-Momentum (Val-Mo) model is a percentile (1-100) ranking of stocks based on recent valuation and momentum characteristics.
Value signals exploit the tendency of stock prices to exhibit mean reversion. Over- and under-valued securities tend to revert to more moderate valuation levels over time. Momentum signals exploit the tendency of trends to continue. Upgrades to analyst estimates and/or recommendations tend to correlate to future upgrades and lead future price moves; past price winners tend to be future winners.
StarMine Val-Mo uniquely captures these signals through a potent blend of four of StarMine's proven stock selection models:
Value is captured via the StarMine Intrinsic Valuation
(IV) and Relative Valuation (RV) models.
Momentum
is captured via the StarMine Analyst Revisions Model
(ARM) and Price Momentum (Price Mo) models.
The combination of global ranks on each of these four inputs results in the overall StarMine Val-Mo score for a security. The headline StarMine Val-Mo score featured in StarMine Professional is the Regional Rank, in which a given stock is ranked against all others in its Region. The Global, Sector, and Industry Rank are also available.
The model assigns each security to one of five distinct regions (North America, Developed Europe, Developed Asia ex-Japan, Japan, and Emerging Markets).
Please note that the document is an approximation of the original document.
StarMine Combined Alpha Model (CAM) White Paper
Synopsis:
The StarMine Combined Alpha Model (CAM) combines all available StarMine alpha models in an optimal, static, linear combination. The weights assigned to each model vary by region. Thus, StarMine CAM takes advantage of the fact that some regions, such as the US and Japan, are more value focused while in Developed Europe and Asia ex-Japan momentum plays a larger role. The model intelligently handles missing data and makes use of all available models for a given security. The StarMine models used in StarMine CAM are Analyst Revisions (ARM), Relative Valuation (RV), Intrinsic Valuation (IV), Price Momentum, Earnings Quality (EQ), Smart Holdings, Insider Filings (US only), and Short Interest (US only).
Please note that the document is an approximation of the original document.
StarMine Combined Credit Risk model (CCR) White Paper
Synopsis:
The Combined Credit Risk Model is StarMine’s best estimate of credit risk that incorporates information from the StarMine Structural, SmartRatios, and Text Mining Credit Risk Models into one final estimate of credit risk.
The Combined model uses a logistic regression framework. In addition, the weights partitioned to the Text Mining model and the other two models are conditioned on the volume of text on a given company, such that the weight on the Text Mining model increases with increasing text volume. The Combined model intelligently handles missing data and will make use of all the component inputs available, but requires only one to score a company.
Please note that the document is an approximation of the original document.
StarMine Text Mining Credit Risk model (TMCR) White Paper
Synopsis:
The StarMine Text Mining Credit Risk model (TMCR) is one component of the StarMine credit risk model suite.
StarMine TMCR assesses the risk in publically traded companies by systematically evaluating the language in Reuters News, StreetEvents conference call transcripts, corporate filings (10-K, 10-Q, and 8-K) and select broker research reports to predict which firms are likely to come under financial distress and which are likely to thrive. It is a percentile ranking (1-100) of stocks, with 100 corresponding to the healthiest companies.
At the core of StarMine TMCR is a classic “bag of words” text mining algorithm. A bag of words text mining algorithm breaks a document into its constituent words and phrases and establishes relationships between the frequencies of these words and phrases and a known training variable, such as observed defaults. Although the StarMine TMCR algorithm and key words and phrases are proprietary, the TMCR components provide insight into the drivers on a stock:
Please note that the document is an approximation of the original document.
StarMine SmartRatios Credit Risk model (SRCR) White Paper
Synopsis:
The StarMine SmartRatios Credit Risk Model is one component of the StarMine Credit Risk Model suite.
The SmartRatios Model is an intuitive and robust default prediction model that provides a view of a firm’s credit condition and financial health by analyzing a wide array of accounting ratios that are predictive of credit risk. The model groups various accounting ratios, along with industry-specific metrics, into 5 components: Profitability, Liquidity, Leverage, Coverage and Growth which are combined in a logistic regression framework. The final default probability is also a function of geographical region.
The SmartRatios model significantly outperforms traditional accounting-based credit models on default prediction such as Altman Z-score and Ohlson O-score. In addition, it can provide incremental value in an equity investment strategy. Finally, it can also serve as a leading indicator of future changes in agency ratings when the SmartRatios rating and the agency rating differ significantly.
The model produces daily updated estimates of the probability of default or bankruptcy within one year for 35,000 companies globally, including Financials. The default probabilities are also mapped to traditional letter ratings and ranked to produce 1-100 percentile scores.
Please note that the document is an approximation of the original document.
StarMine Structural Credit Risk model (SCR) White Paper
Synopsis:
The StarMine Structural Credit Risk Model (SCR) is one component of the StarMine credit risk model suite.
StarMine SCR evaluates the equity market’s view of credit risk via StarMine’s proprietary extension of the structural default prediction framework introduced by Robert Merton that models a company’s equity as a call option on its assets.
In this framework, the probability of default (pd) equates to the probability that the option expires worthless. StarMine SCR produces daily updated estimates of the probability of default or bankruptcy within one year for 35,000 companies globally, including financials. The default probabilities are also mapped to letter ratings and ranked to create 1-100 percentile scores.
Our analysis shows that StarMine SCR is considerably more accurate at predicting defaults than the Altman Z-Score or a basic Merton model, capturing 85% of default events within a 12-month horizon in its bottom quintile of scored companies. In addition to obvious uses for risk management and fixed income security selection, StarMine SCR can also be used to enhance equity selection performance.
Please note that the document is an approximation of the original document.
StarMine Sovereign Risk Model White Paper
Synopsis:
The StarMine ® Sovereign Risk Model (StarMine SR) evaluates a wide array of macroeconomic, market-based, and political data to estimate the probability that a sovereign government will default on its debt. The model produces updated estimates of the annualized probability of default for over 150 countries at six time horizons: one, two, three, Five, seven and ten years. The one-year default probabilities are also ranked to produce 1 to 100 scores and mapped to traditional letter grades
Please note that the document is an approximation of the original document.
StarMine Short Interest model (SI) White Paper
Synopsis:
The Short Interest (SI) model is a percentile (1-100) ranking of stocks based on investor sentiment using the short interest data collected by US exchanges twice per month.
The SI model uses the widely accepted and empirically verified intuition that high levels of short interest reflect negative sentiment on the part of sophisticated investors and are associated with negative future returns. The model is not only a way to identify short candidates, but also identifies long candidates since the absence of short interest shows a lack of negative sentiment. The SI model also intelligently accounts for cost to borrow and arbitrage strategies in its ranking methodology, with 100 representing the highest rank, or a bullish signal.
Our research suggests that arbitrage related-shorts are less informative. The SI model intelligently identifies arbitrage-related shorts and separates them from 'value shorts' that represent directional bets by informed market participants. This improves standalone signal performance and reduces correlation with the suite of other StarMine quant factors.
The final short interest model is conditioned to mitigate the effect of investors leaving short positions to avoid covering dividend payments and investors entering short positions to capitalize on an arbitrage opportunity due to an ongoing merger or acquisition.
Additionally, our research has shown that, while popular, the number of shares in the float is not a reliable proxy for shares available to short. We have found that institutional ownership percentage provides a better proxy for the supply of shares that a bearish investor has to borrow from. As a result, a relatively low institutional ownership percentage indicates a relatively low supply of available shares to short and a high cost to borrow.
The demand factor of short interest percentage and the supply proxy of institutional ownership percentage are the basis for the SI model. A stock with a relatively high short interest (demand) and a relatively low level of institutional ownership (supply) will result in a bearish signal from the SI model, in the absence of any arbitrage-related short activity.
Please note that the document is an approximation of the original document.
StarMine Smart Holdings model (SH) White Paper
Synopsis:
The StarMine SmartHoldings model is a global stock selection model that ranks stocks based on the expected future increase, or decrease, in institutional ownership. The underlying hypothesis of the model is that institutions show high conviction in buying companies with particular characteristics, where these characteristics are fundamental factors such as price/earnings and debt/equity ratios.
At the core of the model is an algorithm that reverse engineers each fund's purchasing profile based on the underlying fundamental factors of the companies a fund is buying. Once we have the purchasing profile of each fund, we compare the fundamental factors of every stock globally to each fund's purchasing profile to determine the alignment between the stock and the fund. We then aggregate the results across all funds and rank each stock based on the collective appeal or demand for it relative to all other stocks by region. The result is a predictive model that accurately predicts which stocks will see an increase or decrease in institutional ownership. A long/short trading strategy based on this ranking produces high Sharpe ratios and annual returns.
Please note that the document is an approximation of the original document.
StarMine Insider Filings model US White Paper
Synopsis:
StarMine's Insider Filings Model US uses insider holdings and insider trades to determine the overall sentiment of insiders towards their company. Insider purchasing reflects bullish sentiment while insider selling reflects bearish sentiment. A score of 1-10 is a bearish indicator and a score of 91-100 is a bullish indicator.
The model considers insider sentiment along three dimensions: agreement among insiders regarding acquiring or disposing of stock ("Net Buyer Ratio"); amount of purchasing ("Buying Depth"); and amount of selling ("Selling Depth"). Each company receives a 1-100 for each of those three components. The final score combines those three components to rank companies on a 1-100 scale. The StarMine Insider Filings Model US currently only scores US companies.
Please note that the document is an approximation of the original document.
