共有14个Excel程序，每个包含不同的风险度量方法和技术，覆盖了常见的VaR模型和算法。

文件无加密，所有细节一目了然，尤其适合初学者了解计算方法和过程，亦可直接拿来套用。里面包含简单说明。

说明：其中两个 (Bootstrap VaR 和 Monte Carlo simulation VaR) 要求安装 Crystal Ball 2000 (或以上)，其他仅仅需要 Excel即可（ 2000版以上)。

 这么好的东西，不能不收点钱。

 

详细内容如下：

Bootstrap VaR: This workbook uses CrystalBall2000 to produce bootstrap simulations from an inputted sample of profit/loss observations, from which a bootstrapped VaR can be inferred.

 Choleski: This workbook illustrates the Choleski decomposition of a 6x6 variance-covariance matrix.

 Component VaR: This workbook estimates the component (or decomposed) VaR using a variance-covariance approach assuming that P/L is distributed as multivariate normal, for chosen confidence level and holding period.

 Cornish-Fisher VaR and ETL: This workbook estimates VaR and expected tail loss (ETL) for near-normally distributed profit/loss, using the Cornish-Fisher adjustment for non-normality.

 ETL estimator: This workbook estimates expected tail loss (ETL) for 4 alternative distributions (normal, Student-t, lognormal, and Gumbel) of profit/loss or return.

 Gumbel VaR: This workbook estimates VaR assuming P/L is Gumbel-distributed. It estimates (and, where appropriate, plots) VaR(s) for a given confidence level and holding period, a range of confidence levels and a given holding period, a given confidence level and a range of holding periods, and ranges of both confidence level and holding period.

 Historical simulation: This workbook estimates and plots historical-simulation VaRs and ETLs for a range of confidence levels.  

Lognormal VaR: This workbook estimates VaR assuming that arithmetic returns are lognormally distributed. It estimates (and, where appropriate, plots) VaR(s) for a given confidence level and holding period, a range of confidence levels and a given holding period, a given confidence level and a range of holding periods, and ranges of both confidence level and holding period.

 Monte Carlo simulation VaR: This workbook estimates normal VaR and the VaR of a long European call option by Monte Carlo simulation.

 Normal VaR: This workbook estimates VaR assuming that P/L and/or arithmetic returns are normally distributed. It estimates VaR and ETL for a given confidence level and holding period, and estimates and plots VaRs for a range of confidence levels and a given holding period, a given confidence level and a range of holding periods, and ranges of both confidence level and holding period.

 Options VaR: This workbook estimates the VaR of vanilla Black-Scholes options using analytical methods.

 t VaR: This workbook estimates VaR assuming that P/L is Student-t distributed. It estimates (and, where appropriate, plots) VaR(s) for a given confidence level and holding period, a range of confidence levels and a given holding period, a given confidence level and a range of holding periods, and ranges of both confidence level and holding period.

 Variance-covariance VaR: This workbook uses a varianace-covariance approach to estimate the VaR of a portfolio assuming that returns are distributed as multivariate normal.

Backtest: This workbook computes an illustrative Kupiec backtest.