The Effect of SFAS No. 131 on the Diversifiction Discount Seoungpil Ahn Sogng Business School, Sogng University PA706, 35 Bekbeom-ro, Mpo-gu, Seoul 121-742, Kore E-mil: sphn@sogng.c.kr Received: July 2, 2015 Accepted: August 1, 2015 Published: December 1, 2015 doi:10.5296/f.v7i2.7952 URL: http://dx.doi.org/10.5296/f.v7i2.7952 Abstrct For smple of diversified firms, I investigte the impct of the segment reporting rule chnge from SFAS No. 14 to SFAS No. 131 in 1997. This chnge in segment-reporting rules to SFAS No. 131 potentilly llows more precise estimtion of diversifiction discount. I probe the chnges in the diversifiction discount before nd fter the reporting rule chnge in 1997. I find tht there is substntil increse in the diversifiction discount under SFAS No. 131. Further nlysis indictes tht the chnges in the diversifiction discount re unrelted to the chnges in firm vlue or investment efficiency. Insted, the mesures of diversity pper to be more ssocited with the chnges in excess vlue. This indictes tht excess vlue is not clen mesure of diversifiction discount. Keywords: Diversifiction; Segment Reporting; Vlution; Diversifiction Discount; Disclosure 58
1. Introduction The issue of corporte diversifiction on vlue remins controversil nd is yet not well understood despite numerous written reports. A prt of the reson for the observed discount in diversified firms ttributed to the errors in segment report. Grhm, Lemmon, nd Wolf (2002), Villlong (2004), nd Whited (2001) ddresses severl concerns in mesuring diversifiction discount with COMPUSTAT segment reportings bsed on industril segments under SFAS No. 14. In June 1997, the Finncil Accounting Stndrds Bord (FASB) issued Sttement of Finncil Accounting Stndrds (SFAS) No. 131, Disclosure bout Segments of n Enterprise nd Relted Informtion, which becme effective for fiscl yers strting fter December 31, 1997. SFAS No. 131 requires enterprises to cover dt bout operting segments, products nd services, the geogrphic regions in which they work, nd their mor customers. This new rule replces SFAS No. 14 tht requires firms to disclose geogrphicl nlysis of foreign opertions to the extent tht foreign opertions ccount for more thn 10% of totl opertions. It is significnt to mention tht the FASB hs given mngement wide ltitude for interprettion of the distinction between domestic nd foreign opertions. Likewise, there is no consistent mnner of reporting the countries of foreign sles. Some firms report their foreign sles for regionl groups such s Europe, Asi, or South Americ, while others report for specific countries such s Jpn, Germny, or the U.K. Under SFAS No. 131, segment reporting is more consistent with the orgniztionl structure of the firm nd provides more detiled informtion bout geogrphic segments. This chnge in segment-reporting rules to SFAS No. 131 my llevites these concerns inherent in SFAS No. 14 nd potentilly llows more ccurte estimtion of diversifiction discount. I exmine the chnges in the diversifiction discount before nd fter the reporting rule chnge in 1997. The result shows tht there is significnt increse in the diversifiction discount under SFAS No. 131. Further nlysis indictes tht the increse in the diversifiction discount is not driven by the poor firm vlution in the lter period, nor it is due to the decrese in the investment efficiency. Rther, the increment in the diversity is ssocited with the observed ltertion in the diversifiction discount. Given tht there is no fundmentl chnges in firm vlue nd investment efficiency, the result suggests tht excess vlue does not necessrily indicte vlue loss due to the diversifiction per se. 2. Smple selection nd estimtion of diversifiction discount 2.1 Smple Selection My initil smple consists of the universe of compnies covered by Compustt (including the reserch file) in both firm level nd segment level t ny time over the period 1992 to 2003. The dt ends in yer 2003 to mtch the smple period before nd fter the rule chge nd lso due to the dt vilbility t my institution. I mtched the firm dt with the segment dt. I obtin segment dt from the Compustt Industry Segment (CIS) files over the period of 1992 to 1997. According to the chnge in the segment reporting stndrd from SFAS No. 14 to SFAS No. 131 in the end of yer 1997, I use segment dt from the Compustt Operting Segment (COS) files over the period of 1998 to 2004. The initil smple of the pper is obtined from the COMPUSTAT Geogrphic Segment (CGS) files for 1998. Since 59
firms effectively dopted the new segment rule SFAS 131 in 1998, I use fiscl yer end 1998 for the smple. To void ny distortion effect of relly smll compnies in clculting the excess vlue, I exclude firms with sles revenue less thn $20 million. I lso exclude firms with ny segment operting in the finncil industry (SIC code 6000-6999), those in other regulted utilities (SIC code 4900-4999) nd those tht re listed s Americn Depository Receipts (ADRs). I lso removed firms lcking the needed informtion in clculting Mrket-to-Sles rtio. I define diversified firms s those firms reporting t lest two segments running in different three-figure, stndrd industril clssifiction (SIC) codes. I define focused firms s those firms reporting only one section. I leve out ny multiple segment firms reported segments operting in the sme three-digit SIC codes. i Therefore, by definition, diversified firms in our smple consist of pure conglomertes operting in unrelted industries. For these diversified firms, I eliminte segments entitled corporte, lloction, inter-segment lloction, nd others. These segments re non-economic ctivities representing unllocted mount in segment sles, segment ssets nd segment operting profits. After elimintion of the non-economic segments, I exclude diversified firms in which the sum of segment sles devites by more thn 1% from the totl sles of the firm. Under SFAS No. 141, firm is required to provide two dditionl yers of bck-dte on segment informtion. If the reported segments re converted from the originl dt yer, the bck-dt could be substntilly different from the originl segment dt. Therefore, I used segment dt which hve source yer mtched with dt yer. 2.2 Estimtion of diversifiction discount The excess vlue of diversified firm is defined similr to Lng nd Stulz (1994), Rn, Serves, nd Zingles (2000), nd Villlong (2004). Following Rn, Serves, nd Zingles (2001), I compute the excess vlue s the nturl log of the Tobin s q of the firm divided by its imputed q. Tobin s is proxied by mrket vlue to sles rtio. Imputed q is clculted s the sles weighted sum of the rtio of the mrket vlue of sles for single-segment firms in the sme industry. Specificlly, the excess vlue of the firm is defined s follows: MV S d Sles Bsed Excess Vlue (EV RSZ) = n ss MV = 1 S S S (MV/S)d is the mrket vlue of firm divided by sles. S is the end-of-the-yer sles of segment nd n is the number of segments. The mrket vlue of firm is the book vlue of ssets plus the mrket vlue of common equity minus the sum of the book vlue of common equity nd blnce sheet deferred tx. Imputed vlues re clculted s the sles weighted 60
sum of the rtio of the mrket vlue to sles for single-segment firms in the sme industry. Thus, (MV/S)ss is the medin mrket-to-sles rtio of single segment firms in the sme industry of segment. I mtch ech segment of diversified firm with the industry medin vlue of single segment firms. Industry is defined t the 4-digit SIC level provided tht there re t lest five single-segment firms in the industry. If there re fewer thn five single-segment firms in 4-digit SIC industry, I define industry t the 3-digit SIC level nd then 2-digit SIC level. I winsorize excess vlues s well s ll other finl vribles t the 1st nd 99th percentiles. I lso use the excess vlue clculted ccording to Berger nd Ofek's (1995) sles multiple pproch nd yield similr results. This selection process resulted in finl smple of 8,147 firm-yer observtions nd totl of 22,632 segment-yer observtions. On verge, the diversified firms report 2.8 segments with medin of 3.0 segments. Over the entire smple period, diversified firms re trding t discount: verge (medin) excess vlue is -0.0988 (-0.0942). The purpose of clculting excess vlue is to provide n unbised estimte of the vlue of the conglomerte when the segments operted s if they re independent firms. I mke severl choices bsed on previous findings to minimize the possibility of bised sttistics. First, I mtch segments with industry medin insted of sles or sset weighted men. I prefer to use industry medin over weighted men becuse the medin is less subect to the size effect. It is possible tht the men vlue is driven by lrge nd, presumbly, more successful single segment firms resulting in downwrd bis in the clculted excess vlue. In fct, when I use weighted verges insted of medin vlue of focused firms, the resulting excess vlue is more negtive. Second, the smple of diversified firms confined to conglomerte hving segments operting t lest two different industries nd exclude firms with multiple segments in the sme industry t the three-digit SIC level. Third, I prefer to use segment sles insted of segment ssets or net income. Segment ssets tend to be under-llocted cross segments nd segment net income is subect to ernings mnipultion. 3. SFAS No. 131 nd the diversifiction discount Tble 1 reports our mesure of excess vlue in ech yer from 1992 2003. I confirm significnt diversifiction discount in the overll period. The excess vlue of diversified firms continues to be discounted under the new segment reporting stndrds. The result suggests tht excess vlue clculted under SFAS No. 14 does correctly cpture the discount in diversified firms. The mgnitude of the diversifiction discount is significntly greter fter the chnge in segment reporting stndrds to SFAS No 131. The excess vlue is on verge -0.05 under SFAS No. 14 nd it is -0.14 under SFAS No. 131. The differences in the men nd medin excess vlue in the two periods re sttisticlly significnt t the 1% significnce level. The significnt increse in the diversifiction discount fter yer 1998 could represent the impct of the segment reporting chnge, the chnge in chrcteristics of diversifiction, or both. 61
Tble 1. Excess vlue of diversified firms over time Under SFAS #14 (1992-1997) Under SFAS #131 (1998-2003) Yer N EV RSZ Yer N EV RSZ 1992 618 1993 599 1994 610 1995 635 1996 611 1997 577 Men [Medin] 3,650-0.0191 [-0.0390] 1998 865-0.0449 b 1999 858 [-0.0522] -0.0612 2000 820 [-0.0713] -0.0507 b [-0.0453] 2001 715-0.0754 2002 641 [-0.0668] -0.0489 b [-0.0465] b 2003 598-0.0498 Men [-0.0538] [Medin] 4,497-0.0050 [-0.0474] -0.1127 [-0.1295] -0.1894 [-0.2051] -0.2182 [-0.2165] -0.1174 [-0.1123] -0.2269 [-0.1947] -0.1386 [-0.1341] The number of diversified firms, excess vlue of diversified firms ech yer, nd the percentge of discount firms. EVRSZ is the nturl log of the rtio of the mrket vlue of totl cpitl to sles for the firm divided by tht of sles-weighted portfolio of focused firms operting in the nrrowest SIC industry.,b,c denote significnce t the 1%, 5%, nd 10% levels, respectively. The smple includes 8,147 firm-yer observtions of diversified firms between 1992 nd 2003. Diversified firms re defined s firms with t lest two segments operting in different 3-digit SIC codes. One possible explntion for the increse in the diversifiction discount under SFAS No. 131 is tht the chnges in firms' environment trigger the increse of diversity leding to deeper discount. ii Alterntively, SFAS No. 14 under-identifies segments nd the dopting SFAS No. 131 corrects the under-identifiction problem. To find out these possibilities, I compre chnges in firm vlue in the two stops. I use Mrket-to-Sles rtio to mesure firm vlution. Tble 2. Diversifiction discount nd firm vlue round segment reporting chnge Under SFAS #14 (1992-1997) Under SFAS #131 (1998-2003) F [χ 2 ] MtoS 1.5957 1.8311 *** 36.62 [1.1661] [1.2027] [78.98] *** EV RSZ is the nturl log of the rtio of the mrket vlue of totl cpitl to sles for the firm divided by tht of sles-weighted portfolio of focused firms operting in the nrrowest SIC industry. *, **, *** denote significnce t the 1%, 5%, nd 10% levels, respectively. The smple includes 8,147 firm-yer observtions of diversified firms between 1992 nd 2003. Diversified firms re defined s firms with t lest two segments operting in different 3-digit SIC codes. 62
Tble 2 indictes tht the corresponding chnge in Mrket-to-Sles rtio is n increse from 1.6 to 1.8. This suggests tht the growth in the diversifiction discount is not forced by lower vlution of diversified firms in the lter period. To further nlyze whether the chnge in excess vlue under SFAS No. 131 is driven by the chnges in investment efficiency, I compre investment efficiency in the two periods. I mesure investment efficiency with reltive vlue dded (RVA). Shin nd Stulz (1998) rgue tht investment efficiency is relted to the diversifiction discount. To compute RVA, I weight firm-nd-industry-dusted segment investment by the difference between the industry medin Tobin s q for tht segment nd the sles-weighted verge q for the firm. More precisely, RVA is defined s follows. Reltive Vlue Added by Alloction(RVA) n ss I n I I I S (q q) w = 1 S S = 1 S S = Totl Sles ss where q is the medin Tobin s q for single-segment firms operting in the sme industry. is sles-weighted verge q of the firm. RVA thus mesures the lloction efficiency of corporte resources. In Tble 3, RVA is -0.003 on verge in the period of 1992-1997 under SFAS No. 14 nd it is -0.0055 in the period of 1998-2003 under SFAS No. The difference is sttisticlly not significnt. The outcome indictes tht there re no substntil chnges in investment efficiency in diversified firms in the two stops. Thus, the observed increse in the diversifiction discount in the two periods is not driven by poor investment efficiency in the lter period. This suggests tht the observed chnge in the diversifiction discount under SFAS No. 131 is not driven by the periodic chnge in firm s fundmentl vlue. Tble 3. Investment Efficiency RVA Under SFAS #14 (1992-1997) Under SFAS #131 (1998-2003) F [χ 2 ] -0.0030-0.0055 2.46 [-0.0003] [-0.0002] [0.49] RVA is mesure of investment efficiency. *, **, *** denote significnce t the 1%, 5%, nd 10% levels, respectively. The smple includes 8,147 firm-yer observtions of diversified firms between 1992 nd 2003. Diversified firms re defined s firms with t lest two segments operting in different 3-digit SIC codes. I next exmine the diversity in the two periods. Diversity is mesured by the number of segments (SEGN), sles-bsed Herfindhl index (Herfindhl), stndrd devition of segment q (STD of segment q), nd Rn et l. s mesure of diversity RSZ Diversity). RSZ diversity is resource-weighted growth opportunities mong segments nd it is mesured by the 63
stndrd devition of sles-weighted industry medin q normlized by rithmetic verge of q. Rn et l. (2000) rgue tht if segments themselves cuse the distortion of cpitl lloction, the mesures of diversity shll be negtively ssocited with RVA. I test the chnge in diversity mesures in the two periods. Previous literture indictes tht the diversifiction discount is relted to the mesure of diversity in wy tht higher diversity of diversified firm is ssocited with lower firm vlue. iii In tble 4, the men [medin] number of segments increses from 2.7 [2.0] to 2.9 [3.0] in the lter period. The chnge is sttisticlly significnt t the 1% significnce level. Similrly, stndrd devition of segment q nd the diversity mesure of Rn et l. (2000) show significnt increse in diversity in the lter period. Although I could not observe ny significnt chnge in Herfindhl index, the overll result from tble 4 suggests tht diversity hs incresed in the lter period. This indictes tht the observed increse in the diversifiction discount is ssocited with the increse in the diversity in the lter period. Thus, the diversifiction discount is incresed under SFAS No. 131 due to the increse in the diversity mesures. However, the increse in the diversifiction discount does not reflect the chnges in fundmentls in firm vlue or investment efficiency. Tble 4. Chnges in diversity Under SFAS #14 (1992-1997) Under SFAS #131 (1998-2003) F [χ 2 ] SEGN 2.67 2.86 *** 77.71 [2.00] [3.00] [78.97] *** Herfindhl 0.5686 0.5659 0.50 [0.5432] [0.5334] [1.98] STD of segment q 0.6033 1.0736 *** 247.28 [0.3341] [0.5156] [167.57] *** RSZ Diversity 0.3058 0.3218 *** 9.81 [0.2605] [0.2713] [7.35] *** SEGN is the number of segments in diversified firm. Herfindhl is sles-bsed Herfindhl index. STD of segment q is the stndrd devition of segment Tobin s q. RSZ Diversity is Rn et l. s mesure of the diversity of diversified firm. *, **, *** denote significnce t the 1%, 5%, nd 10% levels, respectively. The smple includes 8,147 firm-yer observtions of diversified firms between 1992 nd 2003. Diversified firms re defined s firms with t lest two segments operting in different 3-digit SIC codes. 4. Conclusion I exmine the impct of the segment reporting rule chnge on the mesure of diversifiction discount, excess vlue. In 1997, SFAS No. 131 rules replce SFAS No. 14. Under SFAS No. 131, firm shll report informtion bout operting segments, products nd services, the geogrphic res in which they operte, nd their mor customers. With detiled informtion on segments, this chnge in segment-reporting rules my llow more ccurte estimtion of diversifiction discount. 64
The results however show tht the time-series vrition of men excess vlue is not reflecting the chnges in firm s fundmentls. Specificlly, the diversifiction discount is not driven by the chnges in firm vlue nor the decrese in the investment efficiency. The mesures of diversity pper to be more ssocited with the chnges in excess vlue. Becuse I count not find ny evidence tht firm vlue nd investment efficiency hs chnged in the two periods, the observed chnges in excess vlue re more likely to be driven by the number of reported segments, stndrd devition of segment q nd Rn et l (2000) s mesure of diversity. This indictes tht excess vlue is not clen mesure of diversifiction discount. References Berger, P., & Ofek E. (1995). Diversifiction s effect on firm vlue, Journl of Finncil Economics, 37, 39-65. http://dx.doi.org/10.1016/0304-405x(94)00798-6 Grhm, J.R., Lemmon, M.L., & Wolf, J. (2002). Does corporte diversifiction destroy vlue?, Journl of Finnce, 57, 695-720. http://dx.doi.org/10.1111/1540-6261.00439 Lng, L., & Stulz, R.M. (1994). Tobin s q, corporte diversifiction, nd firm vlue, Journl of Politicl Economy, 102, 1248-80. http://dx.doi.org/10.1086/261970 Rn, R., Serves H., & Zingles L. (2000). The cost of diversity: The diversifiction discount nd inefficient investment, Journl of Finnce, 55, 35-80. http://dx.doi.org/10.1111/0022-1082.00200 Shin, H., & Stulz, R. M. (1998). Are internl cpitl mrkets efficient? Qurterly Journl of Economics, 113, 531-552. http://dx.doi.org/10.1162/003355398555676 Whited, T.M. (2001). Is it inefficient investment tht cuses the diversifiction discount? Journl of Finnce, 56, 1667-1692. http://dx.doi.org/10.1111/0022-1082.00385 Copyright Disclimer Copyright for this rticle is retined by the uthor(s), with first publiction rights grnted to the ournl. This is n open-ccess rticle distributed under the terms nd conditions of the Cretive Commons Attribution license (http://cretivecommons.org/licenses/by/3.0/). i Min findings of the pper re not sensitive to this definition. I find qulittively similr results if industry is defined t either the two-digit or four-digit SIC level. ii It is uncler though whether the deeper discount represents mistkes mde by diversified firms, simply the result of cquisition of lredy discounted units, or self-selection. iii Berger nd Ofek's (1995), Lng nd Stulz (1994), nd Rn, Serves, nd Zingles (2000) 65