Question: Who uses your work? Answer:
Our research clients are professional money managers and institutional investors including banks, mutual funds, insurance companies, hedge funds, and plan sponsors. Columbine's subscribers range in size from some of the world's largest investors with billions under management, to smaller "boutique" firms.
Question: How can your research help me?
Here are some of the ways Columbine's alpha forecasting models can help you:
- Save time: Use our models to screen a large universe of stocks down to a small list of strong buy candidates. This lets you spend your time and analytical talents on choosing among those issues that have already been determined to be attractive.
- Save resources: Certainly you could create an alpha forecasting model of your own; you may already have put one together. But building a model that can approach the long-term predictive power and stability of the Columbine models requires tremendous time, experience, and programming resources. The "make vs. buy" decision is clear.
- Improve return: In the highly competitive business of investment management small differences in return can have big payoffs. With the Columbine models you have state-of-the art, alpha forecasting tools that are continuously updated and improved to keep them adding maximum value in your portfolios.
Question: Can I pay for Columbine's services with commissions? Answer:
Yes. Although Columbine is not a broker-dealer, we maintain soft-dollar payment relationships with a number of leading third-party brokers, including PCS Securities, BNY, Second Street, and CAPIS.
Question: Do you manage money with your models?
Columbine does not manage assets.
Question: How many stocks do you cover?
As of September 19, 2013 we cover 5,316 US stocks and 31,399 International (non-US) stocks.
Question: Can I obtain your historical rankings for testing?
Yes. We maintain databases that record how each of the Columbine models ranked specific stocks in the past. This historical ranking data is available to our clients at no charge for use in testing and developing new investment strategies. Prospective clients can purchase a license to receive access our ranking history by payment of a one-time fee. If, after testing the data, you decide to purchase the service, the access fee is credited to the cost of your first year's subscription.
Question: What is the source of the input data for your rankings?
Our source for pricing and company fundamental data is Thomson Reuters. The domestic (US) company data is from their Baseline system. International (non-US) company data is from Thomson Reuters' Worldscope database.
Question: What earnings information do you use in your US models?
In most cases our domestic data supplier, Baseline, reports income from continuing operations
(operating earnings) rather than net income. Where necessary, Baseline adjusts its earnings numbers to conform to those of widely followed publications such as Value Line or Standard & Poors. Baseline's audit process corrects for non-recurring items, discontinued operations, extraordinary items, and accounting changes. As a general rule, Baseline's adjustments conform to First Call's treatment of these items, following the lead of what the majority of broker analysts are doing with that particular company, or specific industry rules.
Question: How do you construct your models?
Everyone in this industry has access to essentially the same historical data. Most model builders select input factors and weightings by running some form of regression analysis on this information. Unfortunately, popular methodology applied to shared data inevitably leads to "me too" models (and portfolios). The only way to stand out from the pack is to bring more appropriate and powerful analytic techniques to bear.
We determine which individual return and risk factors will be included in each model, and set the weightings applied to those factors using Columbine's innovative gradient maximization methodology. This process approaches the construction of a forecasting model as an optimization problem; it searches for the optimal combination of factors that will maximize a model's predictive power. In building our models, the goal or objective function of this gradient maximization
optimization differs depending on the model's intended use. For our Component Models the design objective typically is to maximize 1st
decile spread (gross of transactions costs) at institutional holding periods. Our Stock Selection Models are designed to maximize risk-adjusted portfolio return, net of typical transactions costs, based on monthly rebalancing.
Question: What is the Columbine 1500 Universe?
In the simplest terms, the Columbine 1500 Universe
represents our attempt to identify a reasonably sized pool of stocks that are representative of an investable universe for most institutional money managers. Every week we apply the Columbine models to the data on thousands of domestic stocks - too big a list for many users. To deliver a workable collection of institutional "household name" companies that will be familiar and meaningful for the majority of our clients we need to create a subset of our full production universe. This Columbine 1500 Universe
is the tool that lets us offer full coverage without losing sight of the realistic needs of most of our clients.
Definition: We start with the 500 names from the S & P 500 Composite Index. The additional 1,000 companies are selected on three measures:
Dollar trading volume over the past year
We rank every company in the 2,000 biggest names by cap on each of the three factors individually, and then give each company a composite score based on a combination of those three individual measures. The highest scoring 1,000 issues are added to the S & P 500 stocks to arrive at the Columbine 1500.
Updates: We refresh the Columbine 1500 Universe monthly. Any changes to the S & P 500 Index are automatically reflected in the Columbine 1500. In addition, every month we replace any companies that have gone away through M & A activity, or that have changed their characteristics so drastically that they no longer are entitled to a place in the list. The most typical example would be a company that no longer has any analyst coverage. When a company is removed from the list we replace it with the issue with the next highest-ranking selection score outside the Columbine 1500.