Quantitative
Quantitative
Quantitative
net
Quantitative
Methods
2017CFA二级培训项目
讲师:周琪
1-90
www.gfedu.net
周琪
工作职称:金程教育金融研究院CFA/FRM高级培训师
教育背景:中央财经大学国际经济学学士、澳大利亚维多利亚大学金融风
险管理学学士
工作背景:学术功底深厚、培训经验丰富,曾任课AFP、CFP多年,参与教
学研究及授课,现为金程教育CFA/FRM双证培训老师,担任CFA项目学术
研发负责人,对CFA教学产品的研发工作负责,曾亲自参与中国工商银行总
行、中国银行总行、杭州联合银行等CFA、FRM培训项目。累计课时达400
0小时,课程清晰易懂,深受学员欢迎。
服务客户:中国工商银行、中国银行、中国建设银行、杭州联合银行、杭
州银行、国泰君安证券、深圳综合开发研究院、中国CFP标准委员会、太平
洋保险等
主编出版:参与金程CFA项目参考书目的编写工作,包括金程CFA一级中文
Notes等
2-90
www.gfedu.net
4-90
www.gfedu.net
Reading
9
5-90
www.gfedu.net
1. Scatter Plots
2. Covariance and Correlation
Framework 3.
4.
Interpretations of Correlation Coefficients
Significance Test of the Correlation
5. Limitations to Correlation Analysis
6. The Basics of Simple Linear Regression
7. Interpretation of regression coefficients
8. Standard Error of Estimate & Coefficient of
Determination (R2)
9. Analysis of Variance (ANOVA)
10. Regression coefficient confidence interval
11. Hypothesis Testing about the Regression
Coefficient
12. Predicted Value of the Dependent Variable
13. Limitations of Regression Analysis
6-90
www.gfedu.net
Scatter Plots
A scatter plots is a graph that shows the relationship between the
observations for two data series in two dimensions.
7-90
www.gfedu.net
n
Cov( X , Y ) (X
i 1
i X )(Yi Y ) /( n 1)
Cov( X , Y )
Correlation: r
sx s y
Correlation measures the linear relationship between two
random variables
Correlation has no units, ranges from –1 to +1
8-90
www.gfedu.net
9-90
www.gfedu.net
10-90
www.gfedu.net
11-90
www.gfedu.net
Example
The covariance between X and Y is 16. The standard deviation of X is 4
and the standard deviation of Y is 8. The sample size is 20. Test the
significance of the correlation coefficient at the 5% significance level.
Solution :
The sample correlation coefficient r = 16/(4×8) = 0.5. The t-
statistic can be computed as: 20 2
t 0.5 2.45
1 0.25
The critical t-value for α=5%, two-tailed test with df=18 is 2.101.
Since the test statistic of 2.45 is larger than the critical value of
2.101, we have sufficient evidence to reject the null hypothesis. So
we can say that the correlation coefficient between X and Y is
significantly different from zero.
12-90
www.gfedu.net
13-90
www.gfedu.net
14-90
www.gfedu.net
15-90
www.gfedu.net
16-90
www.gfedu.net
Yi b0 b1 X i i , i 1,..., n
Where
Yi = ith observation of the dependent variable, Y
Xi = ith observation of the independent variable, X
b0 = regression intercept term
b1 = regression slope coefficient
εi= the residual for the ith observation (also referred to as the disturbance
term or error term)
17-90
www.gfedu.net
18-90
www.gfedu.net
19-90
www.gfedu.net
Cov( X , Y )
(X i X )(Yi Y )
b1 i 1
n b0 Y b1 X
(X
Var ( X )
i X) 2
i 1
20-90
www.gfedu.net
21-90
www.gfedu.net
22-90
www.gfedu.net
23-90
www.gfedu.net
Y
Yi b0 b1 X i (Yi Yi ) SSE
__
_
(Yi Y ) SST
(Yi Y ) RSS
__
Y
b0
X
24-90
www.gfedu.net
ANOVA Table
ANOVA Table
df SS MSS
Regression k=1 RSS MSR=RSS/k
Error n-2 SSE MSE=SSE/(n-2)
Total n-1 SST -
SSE
Standard error of estimate: SEE n2
MSE
25-90
www.gfedu.net
SSE
SEE MSE
n2
26-90
www.gfedu.net
27-90
www.gfedu.net
Example
An analyst ran a regression and got the following result:
Coefficient t-statistic p-value
Intercept -0.5 -0.91 0.18
Slope 2 12.00 <0.001
28-90
www.gfedu.net
As SEE rises, sb̂ also increases, and the confidence interval widens
1
because SEE measures the variability of the data about the regression
line, and the more variable the data, the less confidence there is in the
regression model to estimate a coefficient.
29-90
www.gfedu.net
30-90
www.gfedu.net
31-90
www.gfedu.net
32-90
www.gfedu.net
Reading
10
33-90
www.gfedu.net
Framework
2. Interpreting the Multiple Regression
Results
3. Hypothesis Testing about the Regression
Coefficient
4. Regression Coefficient F-test
5. Coefficient of Determination (R2)
6. Analysis of Variance (ANOVA)
7. Dummy variables
8. Multiple Regression Assumptions
9. Multiple Regression Assumption
Violations
10. Model Misspecification
11. Qualitative Dependent Variables
34-90
www.gfedu.net
35-90
www.gfedu.net
Each slope coefficient is the estimated change in the dependent variable for
a one unit change in that independent variable, holding the other
independent variables constant. That’s why the slope coefficients in a
multiple regression are sometimes called partial slope coefficient.
36-90
www.gfedu.net
37-90
www.gfedu.net
Dummy variables
To use qualitative variables as independent variables in a regression
38-90
www.gfedu.net
Dummy variables
Interpreting the coefficients
Example: EPSt = b0 + b1Q1t + b2Q2t + b3Q3t + ϵ
EPSt = a quarterly observation of earnings
per share y x1 x2 x3
Q1t =1 if period t is the first quarter, Q1t
EPSt Q1 Q2 Q3
=0 otherwise
EPS09Q4 0 0 0
Q2t =1 if period t is the second quarter, Q2t
=0 otherwise EPS09Q3 0 0 1
39-90
www.gfedu.net
40-90
www.gfedu.net
Adjusted R2
R2 and adjusted R2
R2 by itself may not be a reliable measure of the explanatory power of
the multiple regression model. This is because R2 almost always
increases as variables are added to the model, even if the marginal
contribution of the new variables is not statistically significant.
Function of adjusted R2
SSE n k 1 n 1 2
adjusted R 2 1
SST n 1
1 1 R
n k 1
adjusted R² ≤ R²
adjusted R² may be less than zero
41-90
www.gfedu.net
can be rejected
Reject H0 if p-value<α
Fail to reject H0 if p-value>α
Regression coefficient confidence interval
j
bˆ t s ˆc bj
Estimated regression coefficient ±(critical t-value) (coefficient standard
error)
42-90
www.gfedu.net
43-90
www.gfedu.net
44-90
www.gfedu.net
45-90
www.gfedu.net
46-90
www.gfedu.net
47-90
www.gfedu.net
48-90
www.gfedu.net
49-90
www.gfedu.net
50-90
www.gfedu.net
51-90
www.gfedu.net
Reject H0,
conclude
positive serial Inconclusive Fail to reject null hypothesis of no
correlation positive serial correlation
0 d1 dU
52-90
www.gfedu.net
53-90
www.gfedu.net
54-90
www.gfedu.net
55-90
www.gfedu.net
Model Misspecification
There are three broad categories of model misspecification, or ways in which
the regression model can be specified incorrectly, each with several
subcategories:
1. The functional form can be misspecified.
Important variables are omitted.
Variables should be transformed.
Data is improperly pooled.
2. Time series misspecification. (Explanatory variables are correlated with
the error term in time series models.)
A lagged dependent variable is used as an independent variable
with serially correlated errors.
A function of the dependent variable is used as an independent
variable ("forecasting the past").
Independent variables are measured with error.
3. Other time-series misspecifications that result in nonstationarity.
Effects of the model misspecification: regression coefficients are biased
and/or inconsistent
56-90
www.gfedu.net
57-90
www.gfedu.net
Credit Analysis
Z – score
Z = 1.2 A + 1.4 B + 3.3 C + 0.6 D + 1.0 E
Where:
A = WC / TA
B = RE / TA
C = EBIT / TA
D = MV of Equity / BV of Debt
E = Revenue / TA
If Z<1.8 Bankruptcy
58-90
www.gfedu.net
Reading
11
Time-series analysis
59-90
www.gfedu.net
1. Trend Models
60-90
www.gfedu.net
Trend Models
Linear trend model
yt=b0+b1t+εt
yt
61-90
www.gfedu.net
Trend Models
Log-linear trend model
yt=e(b0+b1t)
Ln(yt ) =b0+b1t+εt
Model the natural log of the series using a linear trend
Use the Durbin Watson statistic to detect autocorrelation
62-90
www.gfedu.net
Trend Models
Factors that Determine Which Model is Best
A linear trend model may be appropriate if the data points appear to be
equally distributed above and below the regression line (inflation rate
data).
A log-linear model may be more appropriate if the data plots with a
non-linear (curved) shape, then the residuals from a linear trend model
will be persistently positive or negative for a period of time (stock
indices and stock prices).
Limitations of Trend Model
Usually the time series data exhibit serial correlation, which means that
the model is not appropriate for the time series, causing inconsistent b0
and b1
The mean and variance of the time series changes over time.
63-90
www.gfedu.net
64-90
www.gfedu.net
xt 2 b0 b1 xt 1
65-90
www.gfedu.net
No autocorrelation
Covariance-stationary series
No Conditional Heteroskedasticity
66-90
www.gfedu.net
Autocorrelation
Autocorrelation in an AR model
Whenever we refer to autocorrelation without qualification, we mean
autocorrelation of time series itself rxt , xt k rather than autocorrelation of
the error term r t , t k .
Detecting autocorrelation in an AR model
Compute the autocorrelations of the residual
t-tests to see whether the residual autocorrelations differ significantly
from 0,
r t , t k -0 r t , t k
t statistics =
Sr 1/ n
If the residual autocorrelations differ significantly from 0, the model is
not correctly specified, so we may need to modify it (e.g. seasonality)
Correction: add lagged values
67-90
www.gfedu.net
Autocorrelation
Seasonality – a special question
Time series shows regular patterns of movement within the year
The seasonal autocorrelation of the residual will differ significantly from
0
We should uses a seasonal lag in an AR model
For example: xt=b0+b1 xt-1+ b2 xt-4+εt
68-90
www.gfedu.net
Example
Suppose we decide to use an autoregressive model with a seasonal lag
because of the seasonal autocorrelation in the previous problem. We
are modeling quarterly data, so we estimate Equation:
(ln Salest – ln Salest–1) = b0 + b1(ln Salest–1 – ln Salest–2) + b2(ln
Salest–4 – ln Salest–5) + εt.
Using the information in Table 1, determine if the model is correctly
specified.
If sales grew by 1 percent last quarter and by 2 percent four
quarters ago, use the model to predict the sales growth for this
quarter.
69-90
www.gfedu.net
Example
Answer
At the 0.05 significance level, with 68 observations and three
parameters, this model has 65 degrees of freedom. The critical
value of the t-statistic needed to reject the null hypothesis is thus
about 2.0. The absolute value of the t-statistic for each
autocorrelation is below 0.60 (less than 2.0), so we cannot reject
the null hypothesis that each autocorrelation is not significantly
different from 0. We have determined that the model is correctly
specified.
If sales grew by 1 percent last quarter and by 2 percent four
quarters ago, then the model predicts that sales growth this
quarter will be 0.0121 – 0.0839 ln(1.01) + 0.6292 ln(1.02) = e0.02372 –
1 = 2.40%.
71-90
www.gfedu.net
Covariance-stationary
Covariance-stationary series
Statistical inference based on OLS estimates for a lagged time series
model assumes that the time series is covariance stationary.
Three conditions for covariance stationary
Constant and finite expected value of the time series
Constant and finite variance of the time series
Constant and finite covariance with leading or lagged values
Stationary in the past does not guarantee stationary in the future
All covariance-stationary time series have a finite mean-reverting level.
72-90
www.gfedu.net
Covariance-stationary
时间序列从长期来看,往往都带有均值回归
Mean reversion 值时,下一个阶段数值会倾向于减小;而当小
大。以自回归
A time series exhibits mean reversion if it AR(1)模型来看, xt b0
has a tendency to move b1 xt 1
towards its mean
b0 b
mean reverting level is: xt
For an AR(1) model, the程,就可以得到均值 。则当xt
1 b1 1
b0
If xt the model predicts that
b0 x t+1 will be lower
b0 than x t,
(1 b1 ) 接近于 ;当 xt 时,AR(1)模型预
b0
1 b1 1 b1
and if xt the model predicts that x t+1 will be higher than x t
(1 b1 )
案例 5-5,Mean-reverting level
Suppose a one-lag autoregressive model by xt b0
16.54 and 0.65 respectively. If currently X is 42.5,
Referenced Answer
b0 16 .54
73-90
Mean Reverting level = 47 .26
1 b1 1 0.65
www.gfedu.net
Covariance-stationary
Instability of regression coefficients
Models estimated with shorter time series are usually more stable than
those with longer time series
74-90
www.gfedu.net
Random Walks
Random walk
Random walk without a drift
Simple random walk: xt =xt-1+εt (b0=0 and b1=1)
The best forecast of xt is xt-1
Random walk with a drift
xt=b0+xt-1+εt (b0≠0, b1=1)
The time series is expected to increase/decrease by a constant
amount
Features
A random walk has an undefined mean reverting level
A time series must have a finite mean reverting level to be covariance
stationary
A random walk, with or without a drift, is not covariance stationary
75-90
www.gfedu.net
77-90
www.gfedu.net
t2 a0 a1 t21 ut
If the coefficient a1 is significantly different from 0, the time series is
ARCH(1), If a time-series model has ARCH(1) errors, then the variance of
the errors in period t + 1 can be predicted in period t.
If ARCH exists,
the standard errors for the regression parameters will not be correct.
Generalized least squares must be used to develop a predictive model.
78-90
www.gfedu.net
79-90
www.gfedu.net
80-90
www.gfedu.net
81-90
www.gfedu.net
Yes No
模
型
组
建 线性趋势 指数趋势 判断是否有季节性因素
线 a linear trend an exponential trend Seasonality?
路
图
使用DW检验判断残差是否自相关
Yes
Serial correlation?
No Yes
使用趋势模型 使用自回归模型
Use a trend model Use an AR model
82-90
www.gfedu.net
以差额法重新组建序列 以AR(1)模型开始
Take First Differences 模型的估计
自
回
归 残差是否自相关 Yes 继续增加自回归数量和级数
模
型 Serial Correlation? Adding Lags
组
建 No
路
线 是否存在季节性因素 Yes 增加相应的自回归级数
图 Seasonality Present Adding Lags
No
用ARCH模型检测是否存在异质性 Yes 通过广义的最小二乘法来调
Heteroskedasticity? 整模型中的错误
No
组建完成模型,测试模型的预测能力
83-90
www.gfedu.net
Reading
12
84-90
www.gfedu.net
1. Simulation
85-90
www.gfedu.net
Simulation
Steps in Simulation
Determine “probabilistic” variables
Define probability distributions for these variables
Historical data
Cross sectional data
Statistical distribution and parameters
Check for correlation across variables
When two variables are strong correlated, one solution is to pick
only one of the two inputs; the other is to build the correlation
explicitly into the simulation.
Run the simulation
86-90
www.gfedu.net
Simulation
Advantage of using simulation in decision making
Better input estimation
A distribution for expected value rather than a point estimate
Simulations with Constraints
Book value constraints
Regulatory capital restrictions
Financial service firms
Negative book value for equity
Earnings and cash flow constraints
Either internally or externally imposed
Market value constraints
Model the effect of distress on expected cash flows and discount
rates.
87-90
www.gfedu.net
Simulation
Issues in using simulation
GIGO
Real data may not fit distributions
Non-stationary distributions
Changing correlation across inputs
88-90
www.gfedu.net
89-90
www.gfedu.net
生命苦短,如果你有一个机会跟随自己的激情去做你认为重要的事,那么这
个机会就是现在。
90-90