Introduction

Bank names have some other indicators when conduct fundamental analysis.

This part mainly conducts two jobs.

• Validate the model adaptability for banks
• Validate the fundamental factors for banks
  • test original factors validity
  • research potential factors

Adaptability test approach

Data

• Spread data
  • logged credit spread data from banks and other financial institutions
• Market and economic data (group A)
  • US T-bond Yield: 3M, 2Y, 5Y, 10Y, 30Y
  • US yield slope: 2Y10Y, 3M10Y
  • Stock Market Index: HSI, HSTECH, SHSZ300, Nasdaq Composite
  • Stock Market Volatility: Indices’ 30D and 90D volatility
  • Stock Market Momentum: Indices’ 10D momentum
  • Eurodollar contract: 4M and 8M
• Company (issuer) data (group B)
  • issuer’s stock price
  • issuer’s stock volatility: 30D and 90D
  • A calculated “default distance”
    • $DD=(1-\frac{Total Debt}{Enterprice Value + Cash})\times \frac{1}{Equity Volatility}$

Moving average model

• We divide the spread into two parts

  • Fundamental driven (mean driven): this part reflects more on company’s credit profile. We will research if the level of risk free rate also affect this part
  • Market (or short term sentiment/momentum) driven: we will use the selected group A factors to fit this level’s spread.

• We propose the following model to fit the series

  $$ \hat{Y_t} = CS_{MA\ NDays}+ \theta_1 \text{Factor 1}{t-m} + ... +\theta_n \text{Factor N}{t-m}+\theta_{n+1}w_t $$

The $m$ refers to time lag, $\hat{Y_t}$ is fitted on a rolling windows basis. Similarly we will use MSE and RMSE to evaluate the fit goodness.