Dynamic weighted Portfolio
Dynamic weighted Portfolio
Red Devil Partners
Joon Seong Choi,
Youngjun Yoo,
Richard Park,
Young Hun Kim (YK)
Overview
Our purpose is to develop a stock selection strategy, including buy and sell, by using the multi-factors model in order to outperform S&P 500. In this analysis, we evaluate several factors being attributable to stock selection and identify several key factors to make our best model. We also establish a dynamic model and then compare fixed weight factors model and dynamic.
Source Data
We tried to make a strategy for South Korea stock market but we could not do because there were not enough data to bring a reliable result. Instead, we chosen S&P 500 because this pool is one of representatives as well as S&P 500 has several alternative products such future or option in order to flourish our alternatives for making better return or easy hedging.
Methodology
We tested our trading strategy from 2000 to 2005 and plan to have the holding period of 1 month. The next first step was to specify list of factors. Then we did a sort of univariate screens through Factset and evaluated all of them for our stock selection along with identifying 5 fractiles. We brought 4 key factors, Cashflow to Price, Debt to Equity, Market Capitalization, and Price to Book, and selected significant significant portfolios. As a result, we were able to a strategy that could outperform S&P 500. Additionally, we did a dynamic weight strategy and compared fixed weight strategy and dynamic weight strategy.
Factors: we lagged one month.
- Cash to Price
We tested Cash free to Price because we reasoned that the company having high cash to price ratio would be a good stock for our return. However, in equal weight, we could not find out any significant relations as it is seen below. In weighted, we chosen the 5th becaue it had –0.34 return and we could use our short strategy selection.
Equal
[pic]
Weighted
[pic]
[pic]
Debt to Equity
We reasoned that Debt to equity would be an important factor. However, in equal weight, we could not find out any significant relations as it is seen below. In weighted, we chosen the 5th because it had –0.43 return and we could use our short strategy selection.
Equal[pic]
Weighted
[pic]
[pic]
Long term growth rate estimate
We tested long term growth rate estimate. However, in the both, we could not find out any significant relations as it is seen below.
Equal
[pic]
Weighted
[pic]
[pic]
Marketcap
Surprisingly, Marketcap had a sort of consistency. In equal, we used the first group with -0.1 and in weighted, the fifth with 2.06.
Equal
[pic]
Weighted
[pic]
[pic]
NI 3 year growth
We tested NI 3 year growth. However, in the both, we could not find out any significant relations as it is seen below.
Equal
[pic]
Weighted
[pic]
[pic]
Price to Book
We could see the significant in equal and would use in our stock selection. However, we could not fine in weighted.
Equal
[pic]
Weighted
[pic]
[pic]
Price trend momentum by using (price 1m ago-price 2m ago)-1
The result was not the same that we expected. There were nothing we could use.
Equal
[pic]
Weighted
[pic]
[pic]
ROE
We tested ROE. However, the result was disappointing. In the both, we could not find out any significant relations as it is seen below.
Equal
[pic]
Weighted
[pic]
[pic]
Actually, we tested many factors such as ratio of Sales revenue change (revenue 1mago – revenue 2m ago)-1, PE ratio change, EPS, EPS change ratio like comparing by month by month, reinvestment, etc. However we could not reliable or meaningful data. In many cases, we faced so many N/As that we could not some factors that we had thought of important.
In this case, we chosen 4 factors that were Cashflow to Price, Debt to Equity, Market Capitalization which was used in both of equal and weighted, and Price to Book as the basis for our selection.
Optimization with fixed weight
[pic]
We did optimization where our target was the same std of S&P 500. Our model of selecting with having 4 key factors and 5 portfolio outperformed 1.32% of monthly return as it is seen above.
Dynamic weight with using dummy variable
We also tried to make a dynamic strategy model. We added a dummy variable of 3 months S&P500 momentum comparing nearest last 3 month average return last month and 2 month ago. If the moment was week, which was a negative momentum, our model would buy more stocks in the portfolio with negative correlation with S&P500, which is Price to book (5). So using this dynamic weight model, we could avoid a deep loss when S&P 500 decreased.
[pic]
Comparison
As seen above in the graph, the dynamic weighted model provided us higher monthly return of 1.56%. The total cumulative return would be around 300%.
Conclusion and Further Thoughts
After analyzing our results, we can see that the stock selection strategy model using financial factors is useful for making better return. Another finding is that the dynamic weighted model is better than the fixed. However, our models are on basis of monthly rebalancing, causing huge transaction cost. If we incorporate transaction cost, the results would be more reliable. One of possible alternatives could be using future or options because S&P 500 is well organized index in terms of using such products. Therefore, further study for transaction cost is needed. Additionally, it is also valuable to study significant factors. We think such further studies are expected to yield more promising results.
-----------------------
[pic]
................
................
In order to avoid copyright disputes, this page is only a partial summary.
To fulfill the demand for quickly locating and searching documents.
It is intelligent file search solution for home and business.
Related searches
- crm dynamic log in
- dynamic capabilities examples
- dynamic capabilities framework
- dynamic capabilities pdf
- dynamic capabilities theory
- dynamic capabilities model
- dynamic capabilities concept
- dynamic capabilities definition
- dynamic capability theory
- dynamic capability
- nike dynamic capabilities
- txwic crm9 dynamic main