Quantitative Stock Selection Strategies Based on Momentum

1
Quantitative Stock Selection Strategies Based on
Momentum

• Presented by
• ICARUS MANAGEMENT GROUP
• Krista Deitemeyer Scott Dieckhaus Ian Enverga
  Jeremy Hamblin

February 27, 2006
2
Outline

• Strategy Overview
• Factor Analysis
• Conclusion

3
Strategy OverviewWhy Momentum?

• Momentum strategy can help satisfy many client
  and portfolio objectives
• Determine which securities to overweight and
  underweight in an existing benchmark
• Use it for a long-short strategy
• Many people in the industry dispute the validity
  of such strategies
• Test those pundits

4
Strategy OverviewUniverse Definition

• US common stock
• Market capitalization between 500 million and 1
  billion (scaled for time)

Hypothesis
These firms may have greater price inefficiencies
than those that have a larger market
capitalization
5
Factor AnalysisFactors Examined

• Factor 1 (1m avg volume 1m price change) /
  3m avg volume
• Factor 2 Price / 3m avg price
• Factor 3 Price / 1m avg price
• Factor 4 1m avg price / 1y avg price
• Factor 5 1m avg price / 3m avg price
• Factor 6 1m avg price / 6m avg price
• Factor 7 3m avg price / 6m avg price
• Factor 8 12m net sales / Year ago 12m net
  sales
• Factor 9 (Price - 1m avg price) / 1m avg price

6
Factor AnalysisAverage Monthly Returns

• A look a the average returns of the top and
  bottom fractiles of each factor shows that four
  of the factors are the most promising

7
Factor AnalysisBenchmark Outperformance

• Two factors had performed well when analyzing
  of benchmark outperformance

8
Factor AnalysisCumulative Returns In Sample

• The cumulative returns for a long/short strategy
  show that Factor 4 outperforms the rest

Factor 4
9
Factor AnalysisFactor 4 Average Fractile
Returns

• Factor 4 1m average price / 1y average price
• Average In-Sample monthly returns for each
  fractile shows strong linear relationship

10
Factor AnalysisFactor 4 - Yearly Returns Heat
Map

• Heat map indicates a long/short strategy would be
  profitable every year, except the first out of
  sample year

11
Factor AnalysisFactor 4 Cumulative Returns

• In-sample returns show a huge return in 1999
• Out-of-sample returns are somewhat inconclusive

Out of Sample
In Sample
12
ConclusionFactor 4

• Pros
• Profitable strategy both in-sample and
  out-of-sample
• Cons
• Monthly turnover of around 80 means trading
  costs are very high
• Significant outperformance during 1999 skews
  results

• Recommendation
• Improve on Strategy before implementation

13
ConclusionMomentum Strategies

• Profitable opportunities do exist but trading
  cost issues need to be overcome
• Further Exploration
• Layer a predictive model for up or down markets,
  then implement the strategies that would perform
  the best based on the prediction
• Look at different universes (e.g. Large cap, all
  stocks, emerging markets)
• Optimize fractile size and rebalancing periods

14
Questions?