衍生性金融商品交易，在全球各地的資本市場、金融市場早已蔚然成
風，而國內則正當起步有待開發。由於國人對此一新興金融商品普遍缺乏
認識，市場上的交易並不活絡，主要多為外商銀行之間、或是外商銀行與
大型國際性企業之間的交易。 本研究主要是以遠東紡織公司第 62 期
公司債為研究標的，探討債券衍生性新金融商品的訂價。首先，揚棄早期
理論文獻上對於利率期限結構型態所做之平坦假設，考慮利率期限結構的
變化，並引用 Hull-White (1990)的 Extended Vasicek 一因子利率期限
結構模型，以分辨長短期利率及長短期債券之價格。其次，引用 Black
(1976)的遠期債券選擇權評價模式、和 Jamshidian (1989)的選擇權評價
模式，以估算建構在此利率 / 債券之上的選擇權價格，包括 Call 和
Put， 並分別加總之而得到遠紡第 62 期公司債隱含之 Floor 和 Cap 的
價格。接著，以利率期限結構為基礎，依利率下限法(Floor Approach)估
計遠紡第 62 期公司債甲、乙二券的價格。同時，除了將理論價格與次級
市場實際報價作一比較， 以明其訂價誤差之外， 亦變動Hull-
White(1990) Extended Vasicek 利率期限結構模型中之參數， 以瞭解參
數的變化對訂價誤差的影響。此外，另以電腦模擬的方式，求取在
0%~25% 的區間內，遠紡第 62 期公司債隱含之 Floor 和 Cap 的價格。
最後，於 Jamshidian (1989) 的選擇權評價模式下， 以一因子及二因子
的敏感性分析， 探討模型參數值對 Floor 和Cap 價格的影響。
綜合本研究實證結果，所獲致之結論如下： (1) 利率期限結構模型中
參數的估計值， 將隨著採樣期間的不同而有所差異，進而影響到後續
Floor 和 Cap 的評價。 (2) 模型參數的變動對訂價誤差的影響有限
， Jamshidian 模式和 Black 模式之間的差異很小。 (3) 在電腦模
擬之下， 以 Black 模式和 Jamshidian 模式所估算出來之 Cap 和
Floor 的價值相當接近。 (4) 於一因子敏感性分析中得知：短期利率
向長期平均水準 μ 調整的速度值，和 Cap 及 Floor 的價值呈現反向
的關係。 短期利率的長期平均水準 μ 值，和 Cap的價值呈現正向的關
係，和 Floor 的價值呈現反向的關係。指標利率的波動性 σ 值，和
Cap 及 Floor 的價值呈現正向的關係。執行利率 χ 值，和 Cap 的價值
呈現反向的關，和 Floor 的價值呈現正向的關係。 在同一個利率重設期
間，隨著重設日之趨近，Cap 和 Floor 價值逐漸增加，於重設日的前一
天達到最大， 且在重設日當天下降到低於前一次利率重設日之價值，如
此週而復始逐次降低以至於債券到期日時為 0。 (5) 於二因子敏感性
分析中得知： (a)當值較小、即短期利率向長期平均水準 μ
調整的速度較慢時，Cap 和Floor 的價值較大，且此時 Cap 和 Floor 的
價值與長期平均水準 μ 值的高低無關。而當  值較大、亦即調整度較
快時，Cap 和 Floor 的價值較小，且此時 Cap 和Floor 的價值與長期平
均水準 μ 值的高低有關。 (b) 一般而言，指標利率的波動性
σ 值，對 Cap 及 Floor 價值的影響為正向。而執行利率 χ 值，對
Floor 價值的影響為正向、對 Cap 價值的影響為反向。但是若短期利率
向長期平均水準 μ 調整的速度較快時(即值較大)，則會抵消波動性
σ 值、及執行利率 χ 值對於 Cap 和 Floor 值的影響作用。故知：短
期利率向長期平均水準 μ調整的速度之快慢，即  值之大小， 對於
Cap 和 Floor 價值之決定，具有關鍵性的影響，頗值得吾人的注意。

While the transactions of the derivative financial
products have been worldwide popular in the capital and
financial market, it just has its beginning here in Taiwan.
Because the public are not familiar with these new financial
products, the transaction in the market is still not active.
Most of the transactions are among foreign banks or between
foreign bank and large international enterprises. This
study focuses on Far-East Textile 62th corporate bond, and
investigates the pricing of bond derivative financial products.
First, I abandon the flat hypothesis on the term stucture of
interest rate in earlier theoretical paper, and take the
variation of term structure of interest rate into consideration.
I also utilize Hull-White(1990) Extended-Vasicek one factor term
structure of interest rate model, so as to distinguish the
short-term and long-term interest rate, and the prices of short-
term and long-term bond. Secondly, I apply Black(1976) forward
bond option pricing model and Jamshidian(1989) option pricing
model in order to estimate the option prices based on these
rate/bond ,including Call and Put, then summing up to calculate
Floor and Cap price implied in Far-East Textile 62th corporate
bond. Next, applying term structure of interest rate and Floor
Approach, I also calculate the price of Far-East Textile 62th
corporate bond A & B. Meanwhile I compare the theoretical price
with the second market''s quote price, so as to demonstrate its
pricing error. By varying the parameter in Hull-White Extended-
Vasicek term structure of interest rate model, We can find out
the effect of parameter''s variation on the pricing error.
Besides, I use the computer simulation to calculate the Floor
and Cap price implied in Far-East Textile 62th corporate bond A
& B, within the 0%~25% interest rate interval. Last, under
Jamshidian(1989) option pricing model, with one-factor and two-
factor sensitivity analysis, I also investigate parameter''s
influences on Floor and Cap price. To Sum up, the
conclusions are as follows: (1)The estimations of
parameters in the term structure of interest rate model will
vary with different sampling periods. It will then influence the
sequential valuation of Floor and Cap. (2)The variation
of model parameter has very little influence on pricing error.
There is little difference between Jamshidian model and Black
model. (3)Under computer simulation, the Cap and Floor
prices estimated by Black model and Jamshidian model are very
close. (4)From one-factor sensativity analysis, We know
that the adjust rate a of the short-term rate toward long-term
average m is negative related to both Cap and Floor prices. The
long-term average m is positive related to Cap price while
negative related to Floor prices. The volatility of index rate s
is positive related to Cap and Floor price. The exercise rate x
is negative related to Cap price,but positive related to Floor
price. Within the same index rate reset period ,as time passes
toward the next reset date, Cap and Floor prices will gradually
increase and come to their maximum on the day just before next
reset date. On the index rate reset date, the prices will
decrease and lower than those on the last reset date. The
process will repeat again and again until the prices come to
zero on the bond expiry date. (5)From two-factor
sensitivity analysis, We know that: (a)When value a is smaller;
i.e. the adjust rate of the short-term rate toward long-term
average m is slower,Cap and Floor prices are bigger. Also, at
this time Cap and Floor prices have nothing to do with the value
of long-term average m. By contrast, when value a is bigger; i.
e. the adjust rate is faster,Cap and Floor prices are smaller,
and at this moment Cap and Floor prices have something to do
with the value of the long-term average m. (b)Generally
speaking, the volatility of index rate s is positive related to
both Cap and Floor prices. The exercise rate x is negative
related to Cap prices but positive related to Floor prices.
However, when the adjust rate of the short-term rate toward the
long-term average m is faster,i.e. as a is bigger,the effect of
volatility s and exercise rate x on Cap and Floor prices will be
weakened accordingly. As a conclusion, the adjust rate a of
short-term rate toward long-term average m, has key influence on
Cap and Floor prices. This should be worthy of our special
attention.