Report Date: 07/23/2010
Copyright © 2007-2010 c4cast.com, Inc.
All rights reserved.
Report Date: 07/23/2010
Copyright © 2007-2010 c4cast.com, Inc.
All rights reserved.
EconomicInvestor database updated: 07/23/2010
Data is updated at the end of each business week.
About Us » Eta Analysis: Behind the Scenes
Summary: The Economy Matters. EconomicInvestor's patented system is able to explain over 90% of current prices for most stocks, mutual funds, and indices, using a consistent set of core macroeconomic variables. Its underlying models of asset prices demonstrate excellent fit to the data and consistency over time. Using a variety of statistics derived from these models, economic risk can be measured and managed in a consistent way, helping improve performance and risk-exposure in different economic conditions by letting managers create more economically robust portfolios.
As opposed to traditional analytical methods, which often explain approximately 5% to 15% of an asset's price variation, EconomicInvestor's patented system is able to account for over 90% of price variation in most stocks, mutual funds, and indices. Not only does it explain a great portion of asset prices, EconomicInvestor uses a consistent set of 18 macroeconomic and financial variables, or factors, in its models (such as the unemployment rate and consumer price index) so that assets are easily comparable to each other both for historical and current data. Its models of asset prices have demonstrated, and continue to demonstrate, excellent fit to the data and consistency over time. Using a variety of statistics derived from these models, EconomicInvestor can measure and manage assets' "economic risk" - risk due to macroeconomic and financial variables - in a consistent way. Additionally, by using current economic information, the models can identify asset classes and individual securities which are currently facing favorable or unfavorable economic conditions. EconomicInvestor's patented tools provide portfolio managers a way to more effectively manage economic risk and create economically robust portfolios, with better performance and risk potential. In short, the system lets managers look through "A Better Window on the Future."
EconomicInvestor's underlying models can be interpreted either as "cointegration equations" or as "reduced form pro forma models," and have demonstrated strong parameter consistency and high in-sample R-squared fit statistics. Statistical processes using data from the 1970s and early 1980s were used to select the factors (and even the number of factors) for the models. After calculating competing models for this data, several candidate sets were chosen and a predictive-failure and out-of-sample predictive-correlation analysis was performed on data from the mid-early 1980s through 1987 (including the crash). This identified EconomicInvestor's final model specifications and its 18 necessary economic factors. Since 1999, when this process was completed, the model specifications have remained constant (except for necessary institutional changes) and continue to perform admirably under rigorous testing conditions in both up-markets and down-markets. Now, EconomicInvestor provides coverage for most seasoned traded assets (over 20,000 stocks, funds, and indices), and portfolios of those assets.
The EconomicInvestor database is updated after market close each week. The base statistical estimates underlying the analysis are based on a rolling regression methodology and using a computationally efficient estimation algorithm. A three year rolling window with daily data (5-day week) is used for model training.
The Eta factors are probability adjusted partial price elasticities. Where a traditional price elasticity is computed with respect to a given percentage change in a covariate, the Eta measure is computed using a constant empirical probability approximately equal to 2.5% (5% two-tail). By computing elasticities with respect to an approximate constant probability value, the Eta Measures are more directly comparable both in terms of scale and also in terms of likelihood of relevance. The approximate two-standard-deviation values used in the Eta Analysis are generally the same level of probability often used in Value at Risk analysis as well as many empirical stress tests. One might recognize the Eta measures as being equivalent to partial Fisherian durations with respect to each of the eighteen MacroRisk factors.
Note that because of the three-year rolling training period, Eta statistics are not reported for unseasoned securities although most issues would still be available for analysis using the return-based Beta+ tools.