Maturity Transformation and Interest Rate Risk in Large European ...

Table 6: Percentage change in the value of an individual loan contract for various shock sizes, maturitiesand methods. Subtable (a) uses the Basel Committee guideline method, (b) uses the simplest flat yieldcurve loan pricing model, (c) uses a model based on Nelson-Siegel forward rates calibrated to historicalparameters, and (d) uses Svensson forward rates also calibrated to historical parameter values. In eachsubtable the rows give the size of the shock, which is a parallel shift in the level of the forward curve,and the columns give the remaining maturity of the loan contract.(a) Basel Committee guideline1 3 10 15 20 30-3.0% 2.1 6.8 19.9 26.8 33.6 39.0-2.0% 1.4 4.5 13.3 17.8 22.4 26.0-1.0% 0.7 2.2 6.6 8.9 11.2 13.0-0.5% 0.4 1.1 3.3 4.5 5.6 6.50.0% -0.0 -0.0 -0.0 -0.0 -0.0 -0.00.5% -0.4 -1.1 -3.3 -4.5 -5.6 -6.51.0% -0.7 -2.2 -6.6 -8.9 -11.2 -13.02.0% -1.4 -4.5 -13.3 -17.8 -22.4 -26.03.0% -2.1 -6.8 -19.9 -26.8 -33.6 -39.0(c) Nelson-Siegel (β 0 , β 1 , β 2 , τ 1 ) = (3.2, −.7, 4.5, 73)%1 3 10 15 20 30-3.0% 1.5 4.4 14.5 21.5 28.2 41.1-2.0% 1.0 3.0 9.6 14.1 18.5 26.7-1.0% 0.5 1.5 4.8 7.0 9.1 13.0-0.5% 0.2 0.7 2.4 3.5 4.5 6.40.0% 0.0 0.0 0.0 0.0 0.0 0.00.5% -0.2 -0.7 -2.3 -3.4 -4.4 -6.21.0% -0.5 -1.5 -4.7 -6.8 -8.8 -12.32.0% -1.0 -2.9 -9.3 -13.4 -17.2 -23.83.0% -1.5 -4.4 -13.8 -19.8 -25.3 -34.7(b) Flat rate δ = 3.2%1 3 10 15 20 30-3.0% 1.5 4.4 14.1 20.5 26.5 37.3-2.0% 1.0 2.9 9.3 13.5 17.4 24.1-1.0% 0.5 1.5 4.6 6.7 8.5 11.7-0.5% 0.2 0.7 2.3 3.3 4.2 5.80.0% 0.0 0.0 0.0 0.0 0.0 0.00.5% -0.2 -0.7 -2.3 -3.3 -4.1 -5.61.0% -0.5 -1.5 -4.5 -6.5 -8.2 -11.12.0% -1.0 -2.9 -9.0 -12.8 -16.1 -21.53.0% -1.5 -4.4 -13.4 -18.9 -23.7 -31.3(d) Svensson (β 0 , β 1 , β 2 , β 3 , τ 1 , τ 2 ) = (2.1, 1.1, 1.8, 5.5, 65, 855)%1 3 10 15 20 30-3.0% 1.5 4.6 18.7 29.2 38.0 49.9-2.0% 1.0 3.1 12.4 19.3 25.1 32.8-1.0% 0.5 1.5 6.2 9.6 12.4 16.1-0.5% 0.2 0.8 3.1 4.8 6.2 8.00.0% 0.0 0.0 0.0 0.0 0.0 0.00.5% -0.2 -0.8 -3.1 -4.7 -6.1 -7.91.0% -0.5 -1.5 -6.1 -9.4 -12.2 -15.62.0% -1.0 -3.1 -12.1 -18.7 -24.1 -30.83.0% -1.5 -4.6 -18.0 -27.9 -35.7 -45.44.2 How good is the Basel Committee interest rate sensitivity guidelinefor loans?4.2.1 Effect of loan-specific pricingThe Basel Committee guidelines outlined in Section 1.1 do not take into account the type of securityfor which an interest rate risk assessment is being performed. In particular, the guidelines may or maynot be suitable for typical loan contracts. We investigate the performance of the Basel Committeeguideline method for interest rate revaluation sensitivity by comparing it to our alternative simpleloan pricing models of Section 3.2 with parametric forward curves specified in 3.3. The results aresummarised in Table 6. Under all methods, positive interest rate shocks (increases in the level of theforward curve) result in decreases in the value of a loan contract with any maturity, and decreasesin interest rates result in revaluation increases. Larger absolute interest rate shocks result in largeabsolute changes in value.The Basel Committee guideline method, like any duration-based sensitivity method 28 , producessensitivities that are symmetric about zero: positive and negative interest rate shocks of equal absolutesize have an equal absolute revaluation effect; in other words, for any given loan maturity, thepercentage decrease in loan price from an interest rate increase is the same as the percentage increasein loan price from an interest rate decrease of equal absolute size. We can see that all the alternativesimple loan models, however, are able to capture the asymmetric revaluation effect of interest ratechanges. At short maturities and for small shock sizes, these asymmetric effects are negligibly small,while at longer maturities and at larger shock sizes we note that loan prices are more sensitive todecreases in interest rates. It is not surprising, therefore, that the Basel Committee guideline methodsometimes overstates, and at other times understates, the interest rate sensitivity of loan prices.From the regulator’s point of view, we might be most concerned with situations where the Basel28 For an introduction to duration-based sensitivity measurement, see Kaufman (1984).21