The History Of Initial Weight Allocation

Published Date: 02 Nov 2017

BUSINESS SCHOOL

BEFM015 Portfolio Management And Asset Allocation

Group 12

MODULE SUPERVISOR:

DR ZHENXU TONG

Student Name Student Number Candidate Number

Zhen ZHANG 620023073 021894

Xige JIAO 620035817 017915

Tiange ZHANG 610054356 032352

Kai ZHENG 610060308 027683

Jianfei SHEN 620029013 014270

Introduction

In this report, we construct the portfolio mainly considering avoiding extra risk, keeping the portfolio more stable and getting a relatively high long term return. The performance we illustrate is from 28th JAN to 19th APR. We make an adjustment based on the performance before 8th FEB. The report mainly includes the portfolio perspective, performance evaluation and attribution. The relative theoretical basis and models applied will be illustrated in whole process of portfolio.

2. IPS

We seek for our target clients as an individual has a stable salary at age 40-60 or a retiree with fixed pension compensation. The fund will invested 100 million USD and holding the stocks for a long time. Our clients should be risk aversion but can burden risk less than 0.01. Moreover, the target return is annual 25% which further outperform the benchmark.

3. Investment strategy

We actively manage the portfolio in order to beat the market. The global-value strategy will applied in long term period. Thus, we attempt to select the large capital & value stocks which are undervalue. Furthermore, to avoid currency risk, large position of American and British stocks will take in account.

We mainly use the top-down approach to select stocks through three stages—analysis the fundaments in global economies and security markets, comparing relative underperform in variable industries and research the nature of companies.

In the aspect of macro economies, we focus on the G13 countries which own mature economic environment with a stable long-term growth perspective. Their equity markets are usually more efficient with high market capitalization, high liquidity and lower volatility. These conditions of equity market were suit for our strategy. Moreover, the rest of world which also achieves our standards could be considered to diversify our portfolio. The fundamentals of 9 countries which we allocated shows as Figure 1.

For micro aspect, we applied CAMP model to evaluate the expected returns in each country compare to the benchmark, S&P Global 1200. The yield of US 10 years government bonds was used as the risk free rate and the relative betas were estimated from their historic 5 years price movements in each country. Then comparing the expected return to the actual return, we accept the countries which are undervalued.

Under the industry analysis, the fund was looking for the suitable industries, which more likely to be undervalued. Since our objective that is establishing a value style portfolio, we should focus on following elements, including P/E, P/B, P/CF, long term growth and dividend yield.

According to the principle we insisted, we prefer industries with lower P/E, P/B, P/CF ratio but reasonable long-term growth and higher dividend yield. As a result, the majority of our stocks are selected from Utilities, Health Care, Financials, Telecoms and Oil & Gas.

In the above table, we combine countries and industries to classify and rank them as A, B and C level. Finally, the majority of our stocks come from the A-level.

According to the Top-down approach, we then try to analyse the factors of company in terms of the following three standards: Compare the relative techniques between the equities and relative benchmark. Compare the current relative techniques with historical position in each stock. Use the Correlation Analysis to reject the unsuitable stocks.

As an example show in following figure, we compared the investment ratios in each stock with their historical date, and in this part, we take RDSA, KO and JNJ as instance to show the three ratios in 2012.

After comparing with the industries level of SP 500, we mainly chose the undervalued stocks, which have low P/E, P/B, P/CF ratio and high dividend yield. These stocks are consistent with the theory of value-style investment strategy. We list the fifteen companies as a partial sample which illustrated above. To sum this stage, the stocks selected were mainly outperformed in past and are still relative undervalued in current market. However, we also select some stocks, such as Brookfield and MasterCard, which do not follow the main strategy and currently own high momentum, to diversify the risk of solo strategy.

4. Initial Weight allocation

After roughly selection, there are nearly 40 stocks in our portfolio pool. In order to diversification, we should exclude the stocks which are highly inter-related before asset allocation. Thus, correlation analysis was considered first.

This is a correlation table in our five years sample; as a result, 30 stocks left and only 4 pairs of stock with high correlation which above 0.7. Then, the optimal weights were model by several methods. Firstly, according to Markowitz’ modern portfolio theory (MPT), the portfolio could be model by using following formula:

BEAM046 Lecture04 Portfolio management.pdf - Adobe Reader

The point of maximum return under variable given risk level which measured by standard deviation could be found, and these points form the efficient frontier of a portfolio. The optimal portfolio should be created with minimum risk and this called global minimum portfolio. If risk-free asset incorporated in modeling, the optimal portfolio could be constructed by maximizing the Sharpe ratio under a given level of expected return. This portfolio should tangent the efficient frontier. However, this optimization based on mean-variance analysis own a core problem which too hard to overcome. The efficient frontier is very instable and need rebalance fluently. This is infeasible in practice when considering transaction cost. Furthermore, this method sometimes produces extreme value especially when short-sell is allowed and risk-free asset could free to borrow. Therefore, we will use the global minimum weights in the first 10 day period as follow:

5. Mid-term Adjustment

After first period, we observed that the short-sale cause the additional volatility and thus bring the downside risk for our investment. Therefore, a new portfolio with better-diversified weights was designed for the following 50 days period.

The Treynor-Black model (1973) was applied for adjustment and addressing the issue of a few stocks dominates in the initial portfolio, such as the McDonald and Coca Cola. In other words, this model could be regards an appreciate methods to strike a balance between aggressive exploitation of perceived security mispricing and diversification. This model implies require return of risk-free since it uses the risk premium of securities. Furthermore, it is also considered the cost of less-than–full diversification comes from non-systematic risk of mispriced stock by measuring the variance of residual and the abnormal return (alpha) compare to benchmark.

This figure explains how the model works and it offsets the loss of diversification and produces higher possible Sharpe ratio. Finally, we obtain the follow weights in each stock and most of stocks are significantly reduce their power of domination.

6. Portfolio Performance

Our group was tracking the portfolio from 28 January 2012 to 19 April 2012, and during this period, our portfolio has increasing 11.75%, which was outperformed the benchmark by 11.29%. As the following charts present, the portfolio was kept increasing in this investment duration. Comparing with the benchmark, the portfolio presents much better performance.

The higher return usually brings higher risk. In the same time, the variance of our portfolio was higher than the benchmark in this tracking duration. Therefore, we try to diversify the unsystematic risk of our portfolio. The following table indicates some details about risk diversified of our portfolio.

In this table, it shows the non-diversifiable risk of portfolio is slightly higher than the risk of benchmark. The diversifiable risk of portfolio is 0.00106%, which is 42.92% of the total risk. Therefore, basic on this result, our portfolio was trying to maximize investors’ interest and minimum the risk. Therefore, our portfolio is mainly achieved the initial objective and investors’ expectation.

According to the following table, we can find the risk-adjusted return of our portfolio is more than the benchmark. Firstly, it indicates the shape ratio, which was adjusted the return by total risk, is higher than the benchmark. The next one is treynor ratio, which measures the return by systematic risk. In the table, we can find this ratio is over 30 times comparing with the benchmark. In addition, the portfolio’s alpha is 0.1789%. The positive alpha ratio indicates the maximum excess return that fund manager can provide from above-average risk-adjusted. Finally, the information ratio is 0.00025, which results the ability of fund manager to obtain returns consistently.

7. Performance Attribution

All securities, which invested in the portfolio, are stocks, in addition, most of them are from the US and UK stocks market. Russell and FTSE indexes consist of large/medium/small-capitalization for value and growth styles appear to be proper as the style benchmark portfolio. In our portfolio, there are also a few stocks from other countries, such as Norway, Brazil, and Australian, etc. We used these stocks to diversify the unsystematic risk of our portfolio. The result is as follows:

According to the stocks we chosen from Russell and FTSE indexes are basic on large capitalization with value styles. Therefore, the chart indicates the performance attribution is mainly provided by this kind of stocks. The portfolio is categorized in large-capitalization value style which as mainly same as our selected objective type.

The Return Decomposition Analysis estimates the sources of additional portfolio returns, which received over the benchmark. Consequently, we are going to follow Spaulding & David’s (2003) idea, the excess return of portfolio should be constituted from allocation effect and selection effect. The formula and result indicate as following:

Allocation Effect = (Portfolio weight - benchmark weight) * (Benchmark return - Benchmark total return)

Selection Effect = Benchmark weight * (Portfolio return - Benchmark return)

Obviously, the selection effect is positive and extremely high in this portfolio. It is demonstrating the capability of particular stocks selection is quite effective. In addition, it also indicates the past data, which was used to construct portfolio, may reflect the current securities’ trend. It is useful in selecting securities. However, the negative allocation effect presents the portfolio’s assets are allocated to several groups slight ineffectively. Therefore, the portfolio’s return is mainly impacted by ability of choosing stocks.

8. Other Issue: Currency Risk

In the portfolio, there are over 30 stocks from 10 different countries. Therefore, the currency exchanges rate will impact the portfolio seriously. According to this situation, the gains & losses in currency exchanges should be considered into transaction. In each stock’s returns, it was obtained in both local currency and USD. The difference between these two returns measure the currency risk that fund manager faces when translating the money back into USD currency. Consequently, to minimum negative impact from currency exchanges, we mainly choose the stocks from US market to avoid the currency risk as possible as we can.