1. |
|
R introduction and practice with VaR analysis
|
R introduction and practice with VaR analysis: R is a language and environment for statistical computing and graphics. It is a GNU project which is similar to the S language. |
|
2. |
|
Sentiment and investment
|
We extend investor sentiment literature and apply it to conduct portfolio allocation tactically in the Korean stock market. We first construct a Korean investors’ sentiment index by considering prior literature and expert opinions. Second, we investigate whether the index can predict both level and cross sectional variations of stock returns. Third, we attempt tactical asset allocation using the index. Our findings correspond to prior literature. The sentiment index constructed predicts both level and cross sectional variations of stock returns. In addition, the tactical asset allocation generates significant excess return after adjusting risks and transaction costs. Our results are useful not only to academic researchers, but also practitioners such as active fund managers, risk managers and traders. |
|
3. |
|
Principal component analysis and regression
|
We design our sentiment index by extracting the first principal component of proxies and defining it as a sentiment index. We use two methods to construct this index. First, we use BW's variables only (Panel A: SENTIMENT with 6 variables). Second, we use all proxies (SENTIMENT with 9 variables). Thus, the sentiment indexes are first principal components of six and nine variables during the data period respectively. Their correlation with proxies is in the sixth column. |
|
4. |
|
Regression
|
Regression examples: 1. Consumption as a function of income, interest rates, wealth, unemployment rate, etc. 2. Earnings as a function of schooling, age, race, sex, etc. 3. Hours worked as a function of children, sex, spouse’s income, wage, etc. |
|
5. |
|
Multiple Regression
|
Multiple Regression; multicollinearity -- If two or more X’s are linearly dependent, then we say they are colinear. |
|
6. |
|
Multiple Regression (2)
|
Multiple Regression (2): In practice, even if there is not perfect colinearity but near perfect colinearity, the same type of problems occur. In particular, as a set of variables
approaches perfect colinearity, the covariance matrix of the estimates of the associated variables explodes. |
|
7. |
|
multicollinearity, more on regression (LR test, Wald test)
|
more on regression (LR test, Wald test): We discuss the methods about jointly testing several regression coefficients |
|
8. |
|
Omitted variable, LR test, Wald test
|
Omitted variable, LR test, Wald test -- How to compare the fit of two models, one of which (the null model) is a special case of the other (the alternative model). What if we do not observe some independent variables |
|
9. |
|
Asymptotics
|
Asymptotics -- if we focus on the fact that T is large, we can make much more headway by using asymptotic distributions as an approximation. |
|
10. |
|
GLS, simultaneous models and 2SLS
|
GLS examples include time series with autocorrelation, heteroskedasticity, Spacial correlation, etc.
4. Random effects models in panel dat |
|
11. |
|
Student investment ideas; 3SLS
|
Student investment ideas; 3SLS -- So more general covariance structures have no effect on the unbiasedness of OLS. They also do not prevent β^hat being a consistent estimator of β. |
|
12. |
|
Student investment ideas; simultaneous equation models
|
In general, assuming we have specified Ω in terms of a small number of parameters, we can estimate those parameters using either maximum likelihood
estimation (MLE) or method of moments (MOM). |
|