financial crisis

We have collected the daily closing price of three benchmark indexes from different countries, specifically US (S&P500), UK (FTSE100) and Hong Kong (HSI). We selected the data sample from Bloomberg and the period is from January 6th 2003 to December 23rd 2010. A portfolio comprised of equal weight in all three assets S&P500, FTSE100 and HSI are computed as we want to equally diversify our risk within these 3 securities. We tested the stationarity of data using Augmented Dickey-Fuller Test in Eviews and the test results are summarised in Table 1.

 

t-statistics

We Will Write a Custom Essay Specifically
For You For Only $13.90/page!


order now

Probability (p-value)

Standard & Poor’s 500 (S&P 500)

-1.719708

0.4211

Financial Times Stock Exchange 100 (FTSE100)

-1.734055

0.4139

Hang Seng Index (HSI)

-1.529055

0.5188

Portfolio

-1.585160

0.4900

Table 1: Augmented Dickey-Fuller Test Statistics for Closing Price

 

 

We are able to conclude that our data, in fact, have a unit root. The data is then transformed into log-return using the formula below and a database of 1906 observations  is computed,

where  represents then closing price of the asset at time . We repeated the test mentioned above to measure the stationarity of the data and the results in summarised in Table 2.

 

t-statistics

Probability (p-value)

Standard & Poor’s 500 (S&P 500)

-35.80342

0.0000

Financial Times Stock Exchange 100 (FTSE100)

-34.18122

0.0000

Hang Seng Index (HSI)

-45.74177

0.0001

Portfolio

-42.93229

0.0000

Table 2: Augmented Dickey-Fuller Test Statistics for log-return

 

 

We proceed with plotting the graph of the log-return against their respective time period in Figure 3. It can be seen that under the time period being studied, there are clear periods of volatility clustering around the year 2008 – 2009 which is when the subprime crisis peaks.

Figure 3: Log of Daily Returns against Time

 

 

The figure above also shows that our series is stationary with most data being centralised around zero. We then utilise the descriptive statistics to analyse the features of our data.

 

S&P500

FTSE100

HSI

Portfolio

Mean

0.000159

0.000212

0.000453

0.000381

Median

0.000820

0.000652

0.000745

0.000726

Maximum

0.109579

0.093843

0.134068

0.105329

Minimum

-0.094697

-0.092656

-0.146954

-0.118384

Std. Dev.

0.013716

0.013049

0.017147

0.014388

Skewness

-0.262713

-0.128083

-0.183161

-0.240604

Kurtosis

13.57033

11.62666

14.20533

13.49777

Jarque-Bera

8895.290
Prob: 0.0000

5915.343
Prob: 0.0000

9982.161
Prob: 0.0000

8770.366
Prob: 0.0000

Table 4: Descriptive Statistics for the Daily Log-returns

 

 

The results in Table 4 shows that the most volatile is the Hang Seng index, with S&P500 and FTSE100 being less risky relatively. This is because the United States and United Kingdom’s economy are well developed compared to Hong Kong. From the Capital Asset Pricing Model (CAPM), investors would normally expect a higher return for investing in HSI index given the systemic risk of the market. This is proven true as HSI provides the highest compensation compared to the other two equity indexes. Using the concept of CAPM, FTSE100 would be the most conservative investment out of all the indexes listed here as it has the lowest volatility and generates a relatively better return than S&P500.

Looking at the skewness and kurtosis of our data, all series exhibit negative excess skewness and are asymmetrical. The skewness of a normal distribution should be zero, negative skewness means that we have a left tail that is relatively long to the right tail. In words, it means that we have observed more daily negative returns than positive returns for the time period that we’ve chosen. The Jarque-Bera test statistics provide a more comprehensive view that all log-return series are not normally distributed at a statistically high significance level of 0.5%.

 

Leave a Reply

Your email address will not be published. Required fields are marked *