Microsoft Word 2007 ichrie conference Proceedings Final-Final 06-06-07. doc

 
2007 Annual International CHRIE Conference & Exposition 
435
Measurement of Interest Rate Exposure 
Based on prior research, I measure interest rate exposure using monthly stock returns. Assuming an 
efficient capital market, exposure can be estimated empirically using stock returns as an aggregate measure that 
captures all relevant information. Changes in interest rates can alter the risk of the cash flow and this information 
should be reflected in the firm’s equity return. The standard empirical research design has been to regress firm 
equity returns on the percentage change in a stock market index and an interest rate or exchange rate (Jorion, 1990; 
Guay, 1998; Hentschel & Kothari, 2001). Since most firms face exposure from the interest rate sensitivity of their 
assets and debt including any off-balance sheet transactions, I measure exposure using the augmented Capital Asset 
Pricing Model (CAPM). In the first stage, I use the following time-series regression of monthly stock returns on the 
Center for Security Prices (CRSP) equal-weighted market index and the percentage change in the 6-month London 
Interbank Offer Rate (Libor) over the 2000-2004 period: 
R
it 
= b
0
+ b
1
R
mt
+ b
2
R
Lt
+ e
it
(1) 
where R
it
is the monthly stock return for firm i in month t, b
0
is a constant, R
mt
is the return on the CRSP 
equal-weighted index for month t, R
Lt
is the percentage change in the 6-month Libor rate in month t, and e
it
is the error term. 
Including market returns in the model explicitly controls for market movements and reduces any 
correlation between the error terms. The 6-month LIBOR is used in the model because it is the benchmark used for 
most floating rate debt. The estimated beta coefficients (b
2
) for each firm measures the percentage change in the rate 
of return on a firm’s common stock against a percentage change in the interest rate. The sign and magnitude of the 
coefficient measures the direction and degree of the firm’s interest rate sensitivity respectively. Once the interest 
rate exposure is estimated, I test the basic relationship between exposure and derivatives and other control variables 
using a pooled cross-sectional regression in which I regress the absolute value of the exposure coefficients are 
regressed on the determinants of exposure: 
| b
2i 
| = 
γ
0
+ 
γ
’
1
Derivatives
i
+ 
γ
‘Controls
i
+ µ
i
(2) 
where | b2i | is the absolute value of i
th
firm’s interest exposure coefficient 
estimated in equation (1), 
γ
0
is a constant, Derivatives is a measure of hedging and Controls is a 
vector of explanatory variables which I describe shortly, u
i
is the error term.
12