The full text on this page is automatically extracted from the file linked above and may contain errors and inconsistencies.
http://clevelandfed.org/research/workpaper/index.cfm Best available copy Worki na P a ~ e r8408 HOLDING COMPANY INTEREST-RATE SENSITIVITY:. BEFORE AND AFTER OCTOBER 1979 by Gary Whalen Working papers of the Federal Reserve Bank of Cleveland are prel i m i nary materi a1s, circulated to stimul ate di scussion and critical comment. The views stated herein are the author's and not necessarily those of the Federal Reserve Bank of Cleveland or of the Board of Governors of the Federal Reserve System. The author gratefully acknowledges research assistance provided by June Gates. December 1 984 Federal Reserve Bank of Cleveland http://clevelandfed.org/research/workpaper/index.cfm Best available copy HOLDING COMPANY INTEREST-RATE SENSITIVITY: BEFORE AND AFTER OCTOBER 1979 Abstract Since October 1979, market i n t e r e s t - r a t e movements have been frequent and large. Over the same time period, f o r a v a r i e t y o f reasons, competition has i n t e n s i f i e d i n both bank l o a n and deposit markets. These developments have changed the b e n e f i t s and costs o f various types o f a s s e t / l i a b i l i t y management s t r a t e g i e s o r a1t e r n a t i v e l y a f i n a n c i a l i n s t i t u t i o n ' s l e v e l o f i n t e r e s t - r a t e r i s k exposure. I n t h i s study, t h e r a t e - s e n s i t i v i t y postures o f a sample o f h o l d i n g companies are examined over the 1977 t o 1983 i n t e r v a l t o determine whether and how a s s e t / l i a b i l i t y management s t r a t e g i e s changed a f t e r October 1979. I n general, the evidence suggests t h a t h o l d i n g companies reduced t h e i r exposure t o r a t e r i s k i n the immediate post-October 1979 period. t h i s change does n o t appear t o have been permanent. However, The data show a reversal o f t h i s p a t t e r n a t a number o f companies i n 1982 and 1983. I. I n t r o d u c t i o n Changes i n market i n t e r e s t r a t e s have been r e l a t i v e l y frequent and l a r g e r e c e n t l y , p a r t i c u l a r l y since October 1979, when t h e Federal Reserve adopted a new procedure f o r monetary c o n t r o l . The new approach placed g r e a t e r emphasis on the supply o f bank reserves and l e s s emphasis on c o n f i n i n g short- term f l u c t u a t i o n s i n the federal funds rate.' As a r e s u l t , the federal fundsrate http://clevelandfed.org/research/workpaper/index.cfm Best available copy and other market rates became re1 atively more volatile. For example, the standard deviation of the quarter-to-quarter change i n the commercial paper r a t e was 79 basis points over the 11-quarter period before the fourth quarter of 1979. Over the ensuing 9 quarters i t increased to 492 basis points. This increase i n variabil i t y has deeply concerned bank managers (and a1so bank analysts, investors, and regulators), because sharp, unanticipated changes i n market r a t e s can produce undesirable changes i n a banking organization ' s net i n t e r e s t margi n and, t h u s , i t s profi tabil i t y and market Val ue. Whether market-rate gyrations adversely affected a particular i n s t i t u t i o n s ' s performance a f t e r October 1979 depends upon the rate- sensitivity posture maintained by the organization d u r i n g this time. 11. The Gap a s an Index of Rate Sensitivity The net i n t e r e s t margin (NIM) impact of a given change i n market rates occurring over some relatively short time period (90 days, f o r example) generally depends upon the type and s i z e of any banking organization's cumul ative rate-sensi t i v i ty gap re1 ative to i t s vol ume of averaging earning asset^.^ This gap i s defined as the difference between the i n s t i t u t i o n ' s volume of rate-sensi t i v e assets (RSAs) and i t s volume of rate-sensi t i v e l i a b i l i t i e s (RSLs). Any asset or l i a b i l i t y t h a t can be repriced a t some time i n the specified interval i s classified a s rate- sensitive and is included i n the respective t o t a l . fol1 ows: CHNIM = $=* AE A CHR, Symbol ical l y , t h i s re1 ationship can be expressed as http://clevelandfed.org/research/workpaper/index.cfm Best available copy where - CHNIM = NIMt - $GAP = RSA NIMt-l, RSL, AEA = t o t a l average earning assets, and CHR = RT - RT-l , where R i s some representative market r a t e o f i n t e r e s t . Given a r e l a t i v e l y s h o r t time horizon, i f an o r g a n i z a t i o n ' s volume o f RSAs exceeds i t s volume o f RSLs, o r i t has a p o s i t i v e gap, changes i n i t s margin should be p o s i t i v e l y c o r r e l a t e d w i t h changes i n market r a t e s over t h a t interval. The reason f o r t h i s r e l a t i o n s h i p i s t h a t more o f t h e i n s t i t u t i o n ' s assets than l i a b i l i t i e s have r a t e s t h a t change as market r a t e s change. So, given r i s i n g market rates, i n t e r e s t income shoul d increase more than i n t e r e s t expense, causing the organization's n e t i n t e r e s t income and N I M t o r i s e as we1 1. The 1arger the gap re1a t i v e t o an i n s t i t u t i o n ' s t o t a l volume o f average earning assets, the 1arger t h e N I M impact o f a given increase i n rates. Obviously, given a p o s i t i v e gap, an organization's N I M f a l l s along w i t h market rates. Conversely, given a r e l a t i v e l y s h o r t time period, changes i n an i n s t i t u t i o n ' s N I M are negatively c o r r e l ated w i t h market- rate changes i f i t has a negative gap (RSLs > RSAs). Again, the 1arger t h e gap r e l a t i v e t o t o t a l average earning assets, t h e l a r g e r the r a t e impact. The N I M o f organizations t h a t have a zero gap o r are balanced (RSAs = RSLs) should n o t vary markedly i n response t o changes i n market rates. 5 It should be noted t h a t t h e so- called t y p i c a l r e l a t i o n s h i p s between bank gap positions, NIMs, and market r a t e s described above may weaken o r even http://clevelandfed.org/research/workpaper/index.cfm Best available copy -4disappear as the hypothesized time horizon i s lengthened.6 That i s , the NIM impact of a r a t e change assumed t o occur over a longer time interval (1 year f o r example) might not be unambiguously re1 ated t o an i n s t i t u t i o n ' s 12-month cumulative gap position. One reason these relationships break down i s t h a t any given cumulative long-term gap position i s consistent w i t h a wide variety of different short-term incremental ( t h a t is, non-cumul ative ) gap positions. The ultimate NIM impact generated by some given change i n r a t e s assumed t o occur over a 12-month period will depend on the distinctive pattern of short-term gaps a t each individual institution. I t will a l s o depend on how the r a t e s on the various types of assets and 1iabil i t i e s already on the i n s t i t u t i o n ' s books respond t o the given change in market r a t e s , on how the short-term gap positions are reshaped over the period as various assets and 1iabil i t i e s mature, and on other factors as well. 7 I I I. Asset/Liabi 1i t y Management Strategy The discussion above suggests t h a t bank management could have elected t o pursue one of two a s s e t / l i a b i l i t y management strategies i n the v o l a t i l e r a t e environment t h a t prevai 1ed a f t e r October 1979. Management could have attempted t o pursue an anticipatory gapping strategy (creating positive gaps prior t o expected r a t e increases and negative gaps prior t o expected r a t e decl i nes 1, or i t coul d have adopted and maintained a zero-gap position during this time. T h e f i r s t strategy implies t h a t management i s willing t o assume more risk t o earn higher expected returns, because anticipatory gapping i s potentially disastrous i f r a t e expectations are not realized. The risks and potential http://clevelandfed.org/research/workpaper/index.cfm Best available copy -5returns from such a strategy depend on the size of the gap opened. A zero-gap strategy implies a choice of lower expected returns i n exchange f o r lower risk. Management's choice of a strategy might be influenced by i t s degree of satisfaction with the organization's NIM a t the outset of any given period, i t s appetite f o r risk, and i t s a b i l i t y t o forecast i n t e r e s t rates.8 Another important factor i s management's abil i ty t o expeditiously a1 t e r the organization's gap position, given a particular r a t e out1 ook. 9 If management i s dissatisfied w i t h i t s organization's NIM level, i f i t has an appetite f o r risk, i f i t forecasts r a t e s with confidence, and i f i t can reshape the organization's bal ance sheet in any desi red fashion, then anticipatory gapping strategy becomes a t t r a c t i v e and i s l i k e l y t o be pursued. On the other hand, i f an i n s t i t u t i o n ' s management i s content w i t h the current margin level , strongly disl i kes taking risks, has 1i t t l e confidence i n i t s a b i l i t y t o forecast rates, and i s unable t o a1 t e r the organization's gap position easily, a zero-gap strategy appears more a t t r a c t i v e . The s h i f t by the Federal Reserve t o a monetary pggregate targeting procedure in October 1979, i n combination w i t h several other forces, radically a1 tered the operating environment of banks (and of a1 1 financial institutions) . These developments affected the potential risks and returns of both kinds of asset11 iabil i t y management strategies and so may have caused management t o reevaluate, and perhaps a l t e r , the strategy previously pursued. In particular, the s h i f t t o a monetary targeting procedure caused both short-term and long-term i n t e r e s t r a t e s t o change more frequently and by much larger amounts than they had i n the past. Irregular unprecedented movements in rates make accurate r a t e forecasting more d i f f i c u l t and anticipatory gapping increasingly risky. http://clevelandfed.org/research/workpaper/index.cfm Best available copy -6A t t h e same time, i n t r a - and i n t e r i n d u s t r y competition were becoming more intense f o r a l l f i r m s supplying f i n a n c i a l services. Regulatory b a r r i e r s t o p r i c i n g and product competition were being eliminated o r circumvented. This increase i n competition p u t pressure on the margins o f banks and a l l other financial institutions. Management might be induced t o gap more aggressively under such circumstances i n an attempt t o delay, o r even reduce, margin shrinkage t h a t stemmed from deregulation. This study represents an attempt t o determine whether and how t h e gapmanagement s t r a t e g i e s pursued by a nonrandom sample o f 41 regional bank holding companies 1ocated i n 11 d i f f e r e n t states changed a f t e r October 1979. IV. Evidence o f Rate Sensi t i v i t v A D i r e c t Measure o f t h e Rate- Sensitivity Gap From 1979 t o 1982, o n l y a l i m i t e d amount o f i n f o r m a t i o n on t h e r a t e s e n s i t i v i t y c h a r a c t e r i s t i c s o f holding company assets and 1i a b i l it i e s was disc1osed i n pub1ished annual reports. - I t i s possible t o construct o n l y 1 gap measure--a year-end, 12-month gap measure--for h o l d i n g companies from a v a i l a b l e data. Even t h i s gap measure requires a judgment about t h e r a t e - sensi ti v i ty characteri s t i cs o f c e r t a i n bal ance-sheet i tems. Thus, the gap measures used here, l i k e any such measures, are r e l a t i v e l y crude indexes o f each company's exposure t o market- rate changes. Examination o f these measures across companies and changes i n these measures over time i n d i c a t e whether and how holding companies a1t e r e d t h e i r rate- sensi t i v i ty postures since 1979. Year-end 1979, 1980, 1981, and 1982 estimated gap f i g u r e s f o r the sample companies are reported i n appendix A. D e t a i l s concerning the construction o f . these measures are i n c l uded i n t h i s appendix as we1 1 The data i n appendix A http://clevelandfed.org/research/workpaper/index.cfm Best available copy i n d i c a t e t h a t o n l y 9, o r 22.0 percent, of the sample companies had p o s i t i v e 12-month gaps at year-end 1979. The mean 1979 gap r a t i o f o r t h e sample companies was -5.9 percent. L i t t l e evidence o f defensive balance-sheet adjustment i s apparent from these gap measures through year-end 1980. negative gaps a c t u a l l y increased t o 36. year-end. The number o f companies w i t h The mean gap h i t -12.8 percent a t The mean absolute value o f the gap rose from 8.3 percent i n 1979 t o 14.4 percent i n 1980, i n d i c a t i n g t h a t the companies generally d i d n o t reduce the s i z e o f t h e i r gap p o s i t i o n (and so t h e i r exposure t o r a t e r i s k ) during 1 980. However, a reversal o f t h e t r e n d toward greater 1i a b i l i t y s e n s i t i v i t y was evident by the end o f 1981. This might r e f l e c t an attempt by banks t o take advantage o f an expected r i s e i n rates. On the other hand, i t might i n d i c a t e a general desire t o move i n t h e d i r e c t i o n o f a zero gap, given t h e unpredictable r a t e movements during t h i s period. gap p o s i t i o n was -8.2 percent. I n t h i s case, the mean 1981 A formal t e s t i n d i c a t e d t h a t t h e d i f f e r e n c e between the 1981 and 1980 mean gap r a t i o s was h i g h l y s i g n i f i c a n t . The change i n the gap was p o s i t i v e a t 33 companies. The data suggest t h a t companies generally reduced t h e i r i n t e r e s t - r a t e r i s k exposure d u r i n g 1981. The absolute value o f t h e gap declined a t 32 o f t h e sample companies, and the mean absolute value o f t h e 1-year gap measures f e l l by roughly 5 percentage p o i n t s t o 9.8 percent. The general movement i n the d i r e c t i o n o f asset s e n s i t i v i t y continued during 1982. Thi r t y - e ig h t companies exhib i t e d p o s i t i v e gap changes. Twenty-one o f the sample companies had p o s i t i v e one-year gaps a t the end o f t h i s year. The mean 1982 gap p o s i t i o n was -0.1 percent. However, the http://clevelandfed.org/research/workpaper/index.cfm Best available copy mediangap was s l i g h t l y p o s i t i v e a t 0.7 percent. The d i f f e r e n c e between the 1982 and 1981 mean gap measures i s again s t a t i s t i c a l l y s i g n i f i c a n t . The absol u t e val ue gap measures i n d i c a t e t h a t h o l ding companies general l y were u n w i l l i n g t o bear as much r a t e r i s k as i n t h e past. The mean absolute . value o f t h e gap again declined t o 6.3 percent and was be1ow t h e 1979 1eve1 The absolute value o f the gaps of 32 companies was lower i n 1982 than i t had been i n 1981. Twenty- five o f the companies reduced t h e absolute value o f t h e i r gaps i n both o f t h e two preceding years. I n d i r e c t Rate- Sensitivity Gap Measures N I M beta. As noted i n section 11, r e l a t i v e l y long- term gap measures (1i k e 12-month measures) provide only 1i m i t e d i n s i g h t on hol ding company exposure t o r a t e changes occurring over shorter i n t e r v a l s , such as a month o r a quarter. Determination o f t h i s exposure requires d e t a i l e d knowledge o f each i n s t i t u t i o n ' s shorter- term gap p o s i t i o n s - - i t s 30- o r 90-day gap. Few h o l d i n g companies published t h e data necessary t o construct such short- term gap measures over the 1979 t o 1982 i n t e r v a l . However, i t i s possible t o obtain two types o f estimates o f h o l d i n g company short- term gap p o s i t i o n s using non-bal ance-sheet data t h a t are a v a i l able. The c o r r e l a t i o n between changes i n an organization's N I M and changes i n market r a t e s occurring over r e l a t i v e l y short time periods generally depends upon i t s short-term gap position. A p o s i t i v e c o r r e l a t i o n i n d i c a t e s i t has a p o s i t i v e short- term gap; a negative c o r r e l a t i o n , a negative short- term gap; a zero c o r r e l a t i o n , a zero short- term gap. This suggests t h a t the regression c o e f f i c i e n t obtained by regressing t h e short- run change i n a hol ding company ' s N I M on the corresponding change i n a representative market r a t e o f i n t e r e s t . http://clevelandfed.org/research/workpaper/index.cfm Best available copy - 9 can be used as an estimate o f i t s short- term gap position.10 t o t h i s c o e f f i c i e n t here as a company's N I M beta. We w i l l r e f e r The sign o f t h e estimated c o e f f i c i e n t i n d i c a t e s t h e nature o f i t s gap--a p o s i t i v e c o e f f i c i e n t , a p o s i t i v e gap and vice-versa. The s t a t i s t i c a l s i g n i f i c a n c e and absol u t e val ue o f t h e c o e f f i c i e n t provide i n s i g h t t o t h e s i z e o f the gap; a s i g n i f i c a n t l a r g e c o e f f i c i e n t imp1i e s a 1arge-gap p o s i t i o n . An i n s i g n i f i c a n t c o e f f i c i e n t suggests t h a t an i n s t i t u t i o n i s roughly balanced. Q u a r t e r l y n e t - i n t e r e s t margin data were used t o estimate such a regression f o r each o f the sample companies f o r several subperiods from t h e f i r s t quarter o f 1977 t o t h e t h i r d quarter o f 1983. The regression r e s u l t s are d e t a i l e d i n appendix B. A1 though a 1arge proportion o f t h e companies had negative 1-year gaps i n 1979, re1a t i v e l y few (16 ) e x h i b i t e d negative regression c o e f f i c i e n t s from the t h i r d quarter o f 1977 t o the t h i r d quarter o f 1979. J u s t two o f these companies had c o e f f i c i e n t s s i g n i f i c a n t a t t h e 10 percent 1eve1 ( 2 - t a i l t e s t ) . Twenty- five o f the companies had p o s i t i v e c o e f f i c i e n t s , suggesting p o s i t i v e short- term gaps. s i gni f i c a n t . However, only two o f these p o s i t i v e c o e f f i c i e n t s were The mean c o e f f i c i e n t f o r t h e companies w i t h negative c o e f f i c i e n t s was -0.0678, was 0.0757. and f o r t h e companies w i t h p o s i t i v e c o e f f i c i e n t s i t The mean absolute value o f the c o e f f i c i e n t s f o r a l l companies was 0.0726 f o r t h e pre-October 1979 period. C o e f f i c i e n t s obtained from regressions estimated over t h e e n t i r e p e r i o d from the f o u r t h quarter o f 1979 t o t h e t h i r d quarter o f 1983 suggest t h a t short- term gap p o s i t i o n adjustments were s i m i l a r t o t h e longer- term gap changes noted above. I n p a r t i c u l a r , a movement i n the d i r e c t i o n o f asset s e n s i t i v i t y i s evident a f t e r 1979. A t o t a l o f 31 companies e x h i b i t http://clevelandfed.org/research/workpaper/index.cfm Best available copy posi tivecoefficients for t h i s interval ; 13 of these are significant. Only ten companies had negative coefficients, w i t h just one being significant. Further, the regression r e s u l t s suggest t h a t companies generally maintained small er short-term gap positions i n the post-October 1979 interval . The coefficient for companies w i t h positive coefficients i s 0.0353; for the companies w i t h negative coefficients, i t i s -0.0219. The absolute value of the coefficient declined a t 33 of the sample companies, and the mean absolute value of the coefficient i s roughly one-half what i t was i n the pre-October 1979 period: 0.0320 as opposed t o 0.0726. However, i f the post-October 1979 period i s broken into two subperiods of roughly equal 1ength (from the fourth quarter of 1979 t o the fourth quarter of 1981 and from the f i r s t quarter of 1982 t o the third quarter of 1983), the regression r e s u l t s suggest short-term rate- sensitivity adjustments not apparent when the entire period i s examined. The results indicate t h a t most companies (36) had positive short-term gaps i n the f i r s t post-October 1979 subperiod. coefficients are significant. Eighteen of the 36 regression This presumably r e f l e c t s the expectation t h a t short-term rates would r i s e over this interval. coefficients i s significant. J u s t one of the five negative The mean coefficient for the companies w i t h positive coefficients was 0.0396, a s opposed t o -0.0382 for the companies w i t h negative coefficients. The mean absolute value of a l l coefficients was 0.0394. Estimated coefficients for the second subperiod suggest the short-term gaps of most companies turned negative toward the end of 1981. T h i s may r e f l e c t deliberate adjustments t o take advantage of an expected decline i n short-term r a t e s or an inabil i ty t o o f f s e t 1iabil i t y composition changes dueto .' ' the introduction of money market deposit accounts (MMDAS) Thi rty- six of http://clevelandfed.org/research/workpaper/index.cfm Best available copy the estimated coefficients are negative for t h i s interval; thirteen of these are significant. The mean of the negative coefficients was -0.0813. of the positive coefficient was 0.0551. coefficients was 0.0782. The mean The mean absol ute value of a1 1 The l a t t e r i s well above the corresponding value f o r the 1979 t o 1981 period, indicating t h a t companies were generally will i n g t o assume more i nterest- rate ri s k a f t e r 1981 Debt index beta. . I t i s possible to derive another measure of i nterest- rate s e n s i t i v i t y f o r pub1 i c l y traded bank holding companies from stock market data. Essenti a1 l y t h i s is accompl i shed by regressing the periodic r a t e of return on a holding company's stock on some type of i nterest- rate index and some broad stock-market index (which has been orthogonalized w i t h respect t o the interest- rate index t o eliminate correl a t i on between the two i ndependent vari abl e s ) 12 . ' A variety of interest- rate indexes have been employed i n previous research. In most studies, the r a t e of return on a debt instrument or bond index has been used; this i s the approach taken i n t h i s study.' A1 though several alternatives were employed, the results reported are from regressions where the r a t e of return on an index of high-grade corporate bonds was used a s the interest- rate index. 14 The estimated coefficient on the bond index return variable i n the regression equation i s an estimate of the market's view of the rate-sensi t i v i t y posture of a holding company. beta i n this study. I t i s termed the -debt index Since bond returns move inversely w i t h i n t e r e s t rates, a positive significant coefficient on the bond return variable indicates t h a t the company's market value decl ines when market rates r i s e . This suggests t h a t the market considers the company to be 1i abi 1i ty-sensi ti ve (RSL > RSA). http://clevelandfed.org/research/workpaper/index.cfm Best available copy Larger positive coefficients suggest larger negative gaps. The market value of companies w i t h relatively large positive gaps should not decline significantly as rates r i s e , because t h e i r profi tabil i ty should move i n tandem w i t h market rates. Thus, such companies should exhibit negative or insignificant positive debt index return coefficients. 15 The regression resul ts for the sample companies appear i n appendix C. Monthly rate-of-return data were used. Again, the regressions are estimated for a variety of subperiods from January 1977 t o September 1983. The mean bond index return coefficient, or mean debt index beta, was 0.0085 f o r the sample companies i n the pre-October 1979 period. Ten of these coefficients are significant a t the 10 percent level i f a 2- tailed hypothesis t e s t i s conducted. The mean coefficient for these 10 companies was 0.0147. The mean debt index beta coefficient was 0.0053 for the sample companies when the regressions were estimated for the e n t i r e post-October 1979 period. The coefficients of 18 companies were lower for t h i s interval than they were i n the preceding period. the l a t t e r period. However, 34 of the coefficients are significant i n Thus, the debt index beta r e s u l t s for the e n t i r e post-October 1979 interval seem t o conflict w i t h the NIM beta results f o r the same period. Regression results for October 1979 t o December 1981 yield a mean debt index beta coefficient of 0.0055 f o r a l l sample companies. The mean coefficient i s 0.0064 for the 29 companies w i t h significant coefficients. These findings suggest t h a t many companies were viewed by the market a s l i a b i l ity- sensitive over t h i s period, a1 though the NIM beta findings, and t o a l e s s e r extent the long-term gap measures, suggest a general movement i n the direction of asset sensitivity. The coefficients of 27 sampl e companies http://clevelandfed.org/research/workpaper/index.cfm Best available copy - 13 were 1ower i n t h i s p e r i o d than they had been before October 1979, confirming t h e decreased w i l l i n g n e s s t o bear r a t e r i s k revealed by t h e o t h e r 2 measures f o r t h i s time period. Results f o r the January 1982 t o September 1983 p e r i o d reveal t h a t t h e mean c o e f f i c i e n t f o r a l l sample companies decl i n e d s l i g h t l y t o 0.0053. However, the mean debt index beta c o e f f i c i e n t f o r 19 companies w i t h s i g n i f i c a n t c o e f f i c i e n t s was 0.0086--above the value f o r s i m i l a r companies i n t h e preceding time period. Thus, the debt index beta r e s u l t s do n o t r e f l e c t t h e marked s h i f t t o short- term 1i a b i l ity s e n s i t i v i t y a f t e r 1981 t h a t i s i n d i c a t e d by the NIM beta measures. The 1982 t o 1983 c o e f f i c i e n t s o f o n l y 18 companies were smaller than the previous period. in The 1982-83 c o e f f i c i e n t s of 27, o r roughly two- thirds, o f the sample companies were below the value f o r t h e pre-October 1979 interval. However, o n l y t e n companies showed c o n s i s t e n t period- to- period declines over t h e e n t i r e 1977-83 i n t e r v a l . These r e s u l t s confirm the bounce-back (suggested by the N I M beta measures) i n t h e w i l l ingness o f h o l d i n g companies t o bear r a t e r i s k . V. A Comparison o f the Findings Obtained Using the A1 t e r n a t i v e Rate-Sensi t i v i ty Measures The d i f f e r e n t measures o f r a t e s e n s i t i v i t y produce s l i g h t l y d i f f e r e n t p i c t u r e s o f changes i n h o l d i n g company gap-management s t r a t e g y from 1977 t o 1983. This p o i n t becomes more c l e a r i f the three d i f f e r e n t r a t e - s e n s i t i v i ty measures derived f o r each company are c o r r e l a t e d w i t h one another across companies f o r each o f the three sub- interval s examined (see t a b l e 1 ). http://clevelandfed.org/research/workpaper/index.cfm Best available copy Tab1e 1 Correlation Coefficients Ratesensitivity measures N I M beta Debt index beta Jan. 1977-Sept. 1979 Oct. 1979-Dec. 1981 Jan. 1982-Sept. 1983 .271 b -.I31 - .294b - .045 N I M beta Jan. 1977-Sept. 1979 Oct. 1979-Dec. 1981 Jan. 1982-Sept. 1983 a. The gap measure used for the October 1979 t o December 1981 period was an average of the 1979, 1980, and 1981 year-end gap figures. For the January 1982 t o September 1983 interval, the gap measure was the 1982 year-end figure. b. Significant a t the 10 percent level, 2-tailed test. http://clevelandfed.org/research/workpaper/index.cfm Best available copy As noted i n section 11, any long-term gap p o s i t i o n can be c o n s i s t e n t w i t h a wide v a r i e t y o f shorter- term gap positions. Thus, t h e r e l a t i o n s h i p between a company's 12-month gap measure and the other r a t e - s e n s i t i v i t y measures i s n o t clear, a p r i o r i . However, since the 12-month gap i s determined by a company's sequence o f shorter- term gaps, i t seems reasonable t o expect t o f i n d a p o s i t i v e c o r r e l a t i o n between the company's 12-month gap and N I M beta, a1though the c o r r e l a t i on m i ght be weak. A s i gni f icant p o s i t i v e c o r r e l a t i o n was discovered, b u t only f o r the October 1979 t o December 1981 i n t e r v a l . The c o r r e l a t i o n s f o r the other i n t e r v a l s were negative and weak. S i m i l a r l y , t h e re1ationship between a company's 12-month gap measure and i t s debt index beta could be loose. However, a negative c o r r e l a t i o n between such measures appears more l i k e l y than a p o s i t i v e one. A s i g n i f i c a n t negative r e l a t i o n s h i p was detected but, as was the case above, only f o r the October 1979 t o December 1981 period. The r e l a t i o n s h i p between the N I M beta and debt index beta measures a l s o i s n o t determinate, b u t a negative r e l a t i o n s h i p appears l i k e l y . A negative c o r r e l a t i o n was found i n only two o f the three periods examined, and none o f the correlations i s significant. Each measure does p a i n t a s l i g h t l y d i f f e r e n t p i c t u r e o f t h e r a t e s e n s i t i v i t y posture o f the sample companies over t h i s time period. However, the three sets o f measures taken together i n d i c a t e t h a t holding companies generally changed t h e i r rate- sensi t i v i ty postures. I P r i o r t o October 1979, the t y p i c a l h o l d i n g company had a negative long-term gap. However, t h e N I M beta and debt index beta measure r e s u l t s suggest t h a t they d i d n o t t y p i c a l l y have l a r g e negative short- term gaps during t h i s time period as well. Changes i n t h e sample companies' 12-month gap and N I M beta measures i n http://clevelandfed.org/research/workpaper/index.cfm Best available copy - 16 the immediate post-October 1979 period suggest t h a t companies reacted t o the rate v o l a t i l i t y i n this interval by moving toward a s s e t sensitivity. The NIM beta r e s u l t s seem t o indicate t h a t most companies managed t o adjust t h e i r short-term gap positions quickly i n this manner. However, the debt index beta results suggest t h a t the market discounted short-term gap adjustments and penalized companies w i t h longer-term negative gaps. The general decline i n the size of a l l of the rate-sensi t i v i t y measures indicates t h a t most companies maintained small e r gap positions d u r i n g t h i s interval . Results for the final subperiod reveal t h a t the rate- sensitivity trends f i r s t evidenced a f t e r October 1979 generally d i d not continue. The NIM beta results indicate t h a t the short-term gaps of many companies turned from positive to negative. Further, two of the three measures suggest e i t h e r t h a t holding companies became more w i l l i n g t o assume interest- rate risk i n the 1982 to 1983 period, or t h a t they were forced t o do so because of an i n a b i l i t y t o o f f s e t changes i n l i a b i l i t y composition t h a t were due t o deposit-rate deregulation. VI. Summary and Conclusions The results suggest t h a t holding companies d i d a1 t e r t h e i r ratesensitivity postures a f t e r October 1979. In the 1980 t o 1981 period, holding companies generally moved toward asset sensitivity and reduced the size of t h e i r gap positions. However, the changes varied across companies and do not appear t o have been permanent. This behavior i s not surprising i n view of the factors influencing management's choice of an appropriate a s s e t / l i a b i l i t y management strategy a s identified above. For example, gapping might have http://clevelandfed.org/research/workpaper/index.cfm Best available copy - 17 appeared l e s s r i s k y (and so more a t t r a c t i v e ) as r a t e v o l a t i l i t y declined i n 1981, and t h e Federal Reserve announced t h a t i t would abandon i t s s t r i c t monetary t a r g e t i n g strategy. On the o t h e r hand, margin pressures may have forced management t o take on more r i s k t o boost expected returns. It i s a l s o possible t h a t h o l d i n g company management became more w i l l i n g t o assume r a t e r i s k i n 1982 and 1983, because i t had f i n a l l y detected and corrected perceived d e f i c i e n c i e s o r improved the asset11 iabi 1ity management p r a c t i c e s used before October 1979. Given t h a t a company's optimal r a t e - s e n s i t i v i t y posture i s a f u n c t i o n o f several f a c t o r s t h a t t y p i c a l l y change over time, i t i s n o t possible t o unambiguously determine a s i n g l e c o r r e c t posture f o r a l l companies f o r a l l times. Thus, i t i s n o t possible t o conclude t h a t the adjustments evident i n the most recent period are inappropriate. This implies t h a t i t would be d i f f i c u l t t o implement a system o f deposit insurance p r i c i n g t h a t t i e s an i n s t i t u t i o n ' s premi um t o a measure o f i t s in t e r e s t - r a t e r i s k w i t h o u t generating a v a r i e t y o f unintended, perhaps undesirable, changes i n bank behavior. U n t i l more i s known about how banks determine t h e i r o v e r a l l r i s k exposure and exposure t o t h e various kinds o f r i s k , the b e n e f i t s and costs produced by changing the incentives f o r banks t o take p a r t i c u l a r types o f r i s k s w i l l remain uncertain. Given t h i s uncertainty, regulatory changes t h a t a f f e c t the w i l l i n g n e s s o f banks t o take r i s k s should be c a r e f u l l y considered. http://clevelandfed.org/research/workpaper/index.cfm Best available copy Appendix A The 12-month gaps l i s t e d i n table 2 were derived by subtracting each i n s t i t u t i o n ' s estimated total volume of rate-sensi tive 1iabil i t i e s ( t h a t is, those subject t o repricing over the ensuing 12-month interval ) from i t s estimated volume of rate- sensitive assets. i n s t i t u t i o n ' s average earning assets. company annual reports. T h i s total was then divided by the A1 1 data were drawn from bank holding Total estimated rate-sensi t i v e 1iabil i t i e s were assumed t o be the sum of 1arge-denomi nation ($100,000) c e r t i f i c a t e s of deposit (CDs), deposits i n foreign offices, federal funds purchased, securities sold under agreement t o repurchase, other debt w i t h an original maturity of one year or 1ess, and 1ong-term debt w i t h a remaining maturity of one year. In addition, the mean r a t i o of money market c e r t i f i c a t e s t o total deposits ( l e s s 1arge CDs) f o r a1 1 banks i n each holding company's s t a t e for each year i n the period was used as an estimate of the percentage of i t s small-denomination time deposits t h a t were rate-sensi tive. ' T h i s percentage times i t s volume of total deposits (1ess 1arge CDs) produced an estimate of rate-sensi t i v e small -denomination time deposits and was i ncl uded i n the 1i abi 1i ty total . Total rate- sensitive assets were the sum of federal funds sold, securities purchased under agreements t o r e s e l l , investment securities w i t h remaining maturity of one year or 1ess, trading account securities, floating- rate 1oans, and fixed- r a t e 1oans with remaining maturi t i e s of one year or less. http://clevelandfed.org/research/workpaper/index.cfm Best available copy Table 2 Company Holding Company 12-Month Gaps Gap 1979 Gap 1980 Gap 1981 Gap 1982 http://clevelandfed.org/research/workpaper/index.cfm Best available copy Tab1 e 2 Hol ding Company 12-Month Gaps (Continued) Company VA-1 VA-2 VA-3 VA-4 VA- 5 Gap 1979 Gap 1980 Gap 1981 Gap 1982 http://clevelandfed.org/research/workpaper/index.cfm Best available copy Appendix B Regression Results: Corn anyCoeff. P t- stat. . NIM Betas Coeff. t- stat. Coeff. t- stat. AL-1 AL-2 AL-3 AL-4 AL-5 .0699 .2314 .0327 ,0307 -.0548 1.38 4.70" 0.55 0.70 -0.82 .0143 .0372 .0107 .0530 0255 0.81 1.72 0.40 2.29* 1.56 .0241 .0547 .0163 .0637 .0376 1.22 2.93* 0.57 2.23" 2.44" MA-1 .0596 1.62 .0150 0.84 .0298 1.92* MO- 1 MO-2 MO-3 MO-4 .I292 -.0104 .0901 -.0300 0.98 -0.25 1.02 -0.66 .0802 -.0170 -.0108 .0152 1.70 -0.75 -0.20 0.90 .I092 .0015 .0195 2.50* -3.82* 0.05 0.86 TN-1 TN-2 TN-3 .0144 .0460 .0613 0.28 0.43 0.79 .0473 -.0027 .0292 1.87" -0.21 2.01* .0662 .0093 .0471 5.70* 0.76 3.09* - .0343 Coeff. t- stat. http://clevelandfed.org/research/workpaper/index.cfm Best available copy Ap~endix B Rearession Resul t s : N I M Betas (Continued) Company Coeff. t- stat. VA-1 VA- 2 VA-3 VA-4 VA-5 1.55 0.05 0.51 2.10" 1.53 - --- * .0869 .0035 -.0222 .0942 .I011 Coeff. - t- stat. Coeff. - Significant a t the 10 percent level , 2- tai 1 t e s t . t- stat. Coeff. - t- stat. http://clevelandfed.org/research/workpaper/index.cfm Best available copy Appendix C Regression Results: Debt Index Betas 82:l Company Coeff. t- stat. Coeff. t- stat. Coeff. t- stat. Coeff. -83: 9 t- stat. .0053 .0068 .0057 .0077 .0108 2.07" 1.29 1.74" 1.96" 2.78" .0080 .W95 2.36" 1.82" . 01 00 -.0010 .0041 .0092 .0080 2.45" -0.26 1.09 2.88" 2.25" .0099 1.93" .0063 2.11" .0003 .0069 .0030 .0042 0.34 3.18" 0.68 -1.03 .0139 .0063 .0079 .0070 .0103 1.87" 1.48 1.91" 1.92" 2.82" .0061 .0046 .0010 .W92 1.52 0.85 0.21 3.02" .0033 .0054 .003 9 0.83 1.09 0.86 ,0008 .0071 .0020 .0067 0.15 1.54 0.41 1.54 - http://clevelandfed.org/research/workpaper/index.cfm Best available copy Appendix C Regression Resul t s : Debt Index Betas (Continued) Corn any Coeff. t- stat. P - - jSi gni f i c a n t - Coeff. t- stat. Coeff. - - a t t h e 10 percent 1eve1, 2- tai 1 t e s t . t- stat. Coeff. t- stat. http://clevelandfed.org/research/workpaper/index.cfm Best available copy Footnotes 1. Although the procedure was changed again i n August 1982, t h e emphasis on reserve supply remains re1a t i v e l y greater than before 1979. 2. Bank non- interest income i s r e l a t i v e l y stable and l e s s than operating expense. Thus, n e t i n t e r e s t earnings are the key determinant o f o v e r a l l p r o fit a b i 1ity . 3. A number o f authors have made a strong case i n favor o f using the concept o f duration analysis t o create an a l t e r n a t i v e index measure o f r a t e s e n s i t i v i t y - - t h e so- called duration gap. For a discussion o f duration-gap model s, see Toevs and Haney ( 1984). While duration- gap measures have a number o f a t t r a c t i v e p r o p e r t i e s r e l a t i v e t o p e r i o d i c gap models, they do have one p a r t i c u l a r l y n e t t 1esome drawback--1 arge amounts o f very d e t a i 1ed a s s e t / l i a b i l i t y c h a r a c t e r i s t i c information are r e q u i r e d t o construct them. This i s why such measures are n o t used i n t h i s study. 4. Actually, precise measurement o f r a t e - s e n s i t i v e assets and l i a b i l i t y t o t a l s i s q u i t e complicated. For example, i n t e r e s t - r a t e and p r i n c i p a l payments received must be considered, prepayments and d e f a u l t s should be estimated, and estimates o f the r a t e - s e n s i t i v e p o r t i o n s o f l i a b i l i t i e s without e x p l i c i t m a t u r i t i e s must be obtained. For a discussion o f these issues, see t h e studies i n footnote 6. 5. This discussion and a l l t h a t follows presume t h a t banks do n o t hedge exposed gap p o s i t i o n s w i t h off- balance- sheet devices such as i n t e r e s t - r a t e f u t u r e s o r other techniques such as i n t e r e s t - r a t e swaps. Available evidence suggests t h a t most banks d i d n o t a c t i v e l y hedge t h e i r gap p o s i t i o n s i n t h i s way from 1979 t o 1982. 6. For an extensive discussion o f problems and complications involved i n gap model s, see Binder and L i ndqui s t ( 1982 ), Kaufman (1 984), and especi a1l y Toevs (1983) and Toevs and Haney (1 984). 7. Again, off- balance hedging could a l t e r these r e l a t i o n s h i p s and i s assumed t o be immaterial. 8. Management's choice o f i n t e r e s t - r a t e r i s k exposure has a d i r e c t and i n d i r e c t impact on the organization ' s t o t a l r i s k exposure, because i t influences other dimensions o f r i s k . For example, i n t e r e s t - r a t e r i s k exposure a f f e c t s an i n s t i t u t i o n ' s c r e d i t r i s k and 1i q u i d i t y r i s k . 9. Both s t r a t e g i e s presume t h a t bank management i s able t o exercise a considerable amount o f c o n t r o l over i t s organization's rate- sensi t i v i t y posture. Realis t i c a l l y, desired bal ance-sheet adjustments take time and can be c o s t l y . Desired gap adjustments may be constrained by c o n f l i c t i n g customer preferences and competi t i v e pressures. http://clevelandfed.org/research/workpaper/index.cfm Best available copy 10. Such a technique was used i n Olson and Simonson (1982). Here, t h e 90-day negotiable CD r a t e was used as the representative market rate. 11. The MMDA was e s s e n t i a l l y t h e f i r s t r e t a i l deposit product without a r a t e c e i l i n g . Financial i n s t i t u t i o n s could thus attempt t o b i d funds away from competitors. However, since t h i s was n o t possible i n the past, customer and competitor r e a c t i o n t o MMDA p r i c i n g d i f f e r e n t i a l s were unknown. As a r e s u l t , p r e d i c t i o n o f i n f l o w s i n t o MMDAs was subject t o e r r o r , and i n f l o w s probably surprised asset/l i a b i l ity managers a t most i n s t i t u t i o n s . MMDAs a t commercial banks went from zero i n November 1982 t o over $185 b i l l i o n by the end o f t h e f i r s t quarter o f 1983. Because the maximum nominal m a t u r i t y on MMDAs i s one month (and the e f f e c t i v e m a t u r i t y could be l e s s ) , these funds c o n s t i t u t e re1a t i v e l y short- term 1i a b i 1iti es. Large i n f l o w s may have resul t e d i n undesired increases i n 1i a b i l ity s e n s i t i v i t y . 12. See Chance and Lane (19801, Lloyd and Shick (1977), Lynge and Zumwal t (1 980), and F l annery and James (1 984b). 13. Actually, the i n t e r e s t - r a t e index should be a measure o f unanticipated r a t e movements--that i s , a whi te- noise process. Formal s t a t i s t i c a l t e s t s i n d i c a t e d t h a t the index used i n t h i s study could be t r e a t e d as a white- noise process and so t h e r a t e series was n o t transformed i n any way. Flannery and James (1984b) found t h a t t h e i r r e s u l t s were n o t a f f e c t e d when they used various o r i g i n a l r a t e series instead o f a pre-whitened series. 14. S p e c i f i c a l l y , t h e Salomon Brothers r a t e o f r e t u r n index f o r a p o r t f o l i o o f high-grade corporate bonds was used i n the regressions reported. 15. It i s uncertain whether asset- sensitive companies w i l l e x h i b i t negative s i g n i f i c a n t c o e f f i c i e n t s . Some observers have argued t h a t the market values o f asset- sensitive companies w i l l n o t change s i g n i f i c a n t l y as market r a t e s change, because the n e t income o f such companies w i l l r i s e and f a l l i n tandem w i t h market r a t e s and, presumably, t h e r a t e s i n v e s t o r s use t o discount t h e i r cash f l o w streams o f banking organizations. http://clevelandfed.org/research/workpaper/index.cfm Best available copy References Aspi nwel 1, Richard C. "Managing I n t e r e s t Rate Risk: The Basic Concepts ," Proceedings o f a Conference on Bank S t r u c t u r e and Competition, Federal Reserve Bank o f Chicago, A p r i l 12 14, 1982 s PP. 463 -70. - Biderman, Mark. " Monitoring and Measuring I n t e r e s t Rate Risks," Proceedings o f a Conference on Bank S t r u c t u r e and Competition, Federal Reserve Bank o f Chicago, A p r i l 12-14, 1982, pp. 411 -83. Binder, B a r r e t t F., and Thomas W.F. Lindquist. A s s e t I L i a b i l i t y and Funds Management a t U.S. Commercial Banks, A Research Report Prepared f o r Bank Administration I n s t i t u t e ' s Accounting and Finance Commission, Roll i n g Meadows, I L : Bank Administration 1 n s t itute, 1982. Chance, Don M., and William R. Lane. "A Re-examination o f I n t e r e s t Rate S e n s i t i v i t y i n the Common Stock o f Financial I n s t i t u t i o n s , " Journal o f Financial Research, vol 3, no. 1 (Spring 1980), pp. 49-55. . Dew, James Kurt. " I n t e r e s t Rate R i sk Management: Maximizing Earnings a t Minimum Risk," Proceedings o f a Conference on Bank s t r u c t u r e and Competition, Federal Reserve Bank o f Chicago, A p r i l 12-14, 1982, pp. Flannery, Mark J., and Christopher M. James. "Market Evidence on t h e M a t u r i t y o f Bank Assets and L i a b i l ities," Journal o f Money, C r e d i t and Banking, vol. 16, no. 4 (November 1984a), pp. 435 45. - . "The E f f e c t o f I n t e r e s t Rate Changes on the Common Stock Returns o f Financial I n s t i t u t i o n s , " Journal o f Finance, vol no. 4 (September 1984b), pp. 1141-53. . 39, Graddy, Duane B., and Adi S. Karna. "Net I n t e r e s t Margin S e n s i t i v i t y among Banks o f D i f f e r e n t Sizes," Journal o f Bank Research, Winter 1984, pp. 283-90. Hanweck, Gerald A., and Thomas E. K i l c o l l i n . "Bank P r o f i t a b i l i t y and - I n t e r e s t Rate Risk," Research Papers i n Banking and ~ i n a n c i a l Economics. D i v i s i o n of Research and S t a t i s t i c s . Board of Governors o f the ~ e d e r a lReserve System, J u l y 1981. Kaufman, George G. "Measuring and Managing I n t e r e s t Rate Risk: Economic Perspectives, Federal Reserve Bank of Chicago, JanuaryIFebruary 1 984, pp. 16-29. A Primer," Lindhart, Ronald A., and Robert R. K linzing. "Measuring and Monitoring I n t e r e s t Rate Risk: A Regulator's Viewpoint," Proceedings of a Conference on Bank Structure and Competi t i o n , Federal Reserve Bank o f 1 12 14, 1982 s PP* 484 91 - - http://clevelandfed.org/research/workpaper/index.cfm Best available copy Lloyd, William P., and Richard A. Shick. "A Test of Stone's Two-Index Model of Returns," Journal of Financial and Q u a n t i t a t i v e Analysis, September 1977, pp. 363-16. Lynge, Morgan J . , J r . , and J . Kenton Zumwal t. "An Empirical Study of the I n t e r e s t Rate Sensi t i v i t v of Commerci a1 Bank Returns: A Mu1 t i -Index Approach, " Journal of ~ i i a n c i a land Q u a n t i t a t i v e Analysis, vol 15, no. 3 (September 19801, pp. 131 42. - . Martin, John D., and Arthur J . Keown. " I n t e r e s t Rate S e n s i t i v i t y and Portfol i o R i sk," Journal of Fi nanci a1 and Q u a n t i t a t i v e Analysis, June 1977, pp. 181-95. Olson, Ronald L., and Dona1 d G. Simonson. "Gap Management and Market Rate S e n s i t i v i t y i n Banks," Journal of Bank Research, Spring 1982, pp. 53-8. Toevs, A1 den L. "Gap Management: Managing I n t e r e s t Rate R i s k i n Banks and ~ h rf it s ," Econorni c Review, Federal Reserve Bank of San Franci sco, no. 2 (Spring v - and William C. Hanev. Measurina and Manaaina I n t e r e s t Rate