Home >> Proprietary Models
A basic tenant of Investment theory is that since a company's current value is equal to the discounted
value of its future earnings, changes in a company's expected future earnings should directly translate
into changes in the company's stock price. In 1982 Zacks developed a model ( the Zacks Indicator )
that predicted future revisions of brokerage analyst EPS estimates using historical patterns in the
brokerage analyst EPS estimates. This was the first "Estimate Revision" model. Over the years since
1982 the Zacks Indicator model has been reformulated a number of times and has proven itself to be perhaps
the most effective technique available to professional investors for predicting price changes over a 3
to 6 month horizon.
During the last 23 years, the portfolio consisting of the top 5% of the companies selected by the Zacks
Indicator delivered an Annualized Return of 35%, compared to 15.9% return for the S&P 500, 13.6% return for
the equal weighted universe and 3% return for the portfolio of the bottom 5%. These returns are not the
result of backtests but are the returns Zacks clients could have achieved (excluding transaction costs)
had they followed the recommendations of the Zacks Indicator as it was delivered to them by
Zacks weekly since 1982. These returns have been confirmed and audited by an independent CPA
firm whose report is available on request.
The Zacks Indicator is a composite score whose ranking identifies stocks whose returns are predicted to be
higher than the equal weighted universe over the next 3 to 6 months. The Zacks Indicator composite score
is calculated for all US and Canadian companies followed by Wall Street analysts ( about 3500 )
and ranks these stocks into 99 equal groupings (1=best and 99=worst). The four components used to
determine the composite scores are:
1. Agreement - Extent to which analysts have been revising their estimates in the same direction.
2. Magnitude - Percent change in the mean consensus estimate.
3. Upside Potential - Extent to which the most accurate estimate deviates from the consensus estimate.
4. Surprise - Pattern of recent deviations of the reported quarterly EPS from the consensus estimates
BUY vs. MAKE?
Since Zacks pioneering work in 1982, databases of estimate revisions have become widely available
from both Zacks and First Call and many large buy side firms have used this data, just as Zacks
first did in 1982, to create their own estimate revision models. In fact a 1999 survey by Merrill Lynch
showed that estimate revision models were even then one of the 3 most widely used models by
institutional investors. The question naturally arises - is the Zacks model better than an in-house model
that you could create? Only you can answer that. We are happy to provide you with the history of the Zacks
Indicator so that you can compare the performance of Zacks vs. your internal model on your specific
IN2 - 2005 INDICATOR FORMULATION
During the last year our team of 5 quantitative analysts (3 of whom have Ph Ds) reformulated the Indicator
to take account of the fact that most large buy side firms are now using some type of estimate revision model.
We added a number of components to the Indicator, relating to changes in analyst recommendations, changes
in analysts sales forecasts, variations in the range of estimates and a more
accurate definition of the Most Accurate Estimate. This new formulation, which we call IN2, was backtested
and dramatically outperformed the Indicator in many investment universes. We believe this performance of
IN2 is due to its ability to predict decisions being made by large buy side firms who are using their own
in- house estimate revision models or using the estimate revision model developed by Starmine in 1999.
As of Jan 2005 IN2 has replaced the Indicator as our primary stock selection model for institutional investors.
BACKTESTING PERFORMANCE OF IN2
Since 1985 Zacks has provided institutional investors with the industry's most comprehensive software system
for designing and testing quantitative equity models. The system, which we call the Zacks Research System was
migrated to the web in 2000 and is available for your use in testing the performance of IN2 within your
investment universe. Contact Adam Cohen, the IN2 product manager at 312-630-9898 x 227 for more information.
ZACKS RANK vs. ZACKS INDICATOR
In 1995 Zacks began providing the Indicator to individual investors in the form of the Zacks Rank, a 1-5 score
assigned to each stock translated directly from the Zacks Indicator. Zacks Rank =1 corresponds to Indicator
from 1 to 5, Zacks Rank =2 corresponds to Indicator from 6 to 20, Zacks Rank =3 corresponds to Indicator
from 21 to 80, Zacks Rank =4 corresponds to Indicator from 81 to 94, and Zacks Rank =5 corresponds to
Indicator from 95 to 99.
MORE INFORMATION ON ZACKS RANK
In 1999 Harper Collings published the book "Ahead of the Market" by Mitch Zacks which detailed how to
use the Zacks Rank in a number of investment processes. This book is available at major book stores
and from your Zacks account manager.
SNAPSHOT REPORT AND ZACKS REC
The Snapshot Report, which is included in our Independent Research offering, is a one page quantitative
research report available for a universe of 1900 companies. The Zacks Buy/Hold/Sell Recommendation at
the top of the Snapshot report is derived from the Zacks Rank as follows: Zacks Rec = Buy consist
of companies whose Zacks Rank was at one time equal to 1 and has remained a 1 or a 2 since that time,
Zacks Rec = Sell consists of companies whose Zacks Rank was at one time equal to a 5 and has remained
a 4 or 5 since that time, all other companies have Zacks Rec = Hold. This definition of the Zacks Rec
was formulated to provide a simple Buy / Hold /Sell recommendation with a lower turnover than the
Zacks Rank. The Zacks Rank and Zacks Rec can be easily calculated directly from the
Zacks Indicator Score.
HOW ESTIMATE REVISION MODELS ARE USED
The Zacks Indicator and Zacks Rank are momentum models whose optimal time horizons are 1-6 months:
consequently many institutional investors use the model for a timing tool rather than a method of
picking long term out performers. For instance, the more common uses of Zacks models include:
One component multi-factor models
Screening for new purchase candidates
Tiebreakers to pick among equally attractive stocks
Tiebreakers to pick among stocks in an industry group