To what extent does the CAPM (efficient capital markets) model describe British capital flows over the
period 18 ! " 1#1$% &t is possible to b'ild a pricin( model for capital assets that (ives the e)'ilibri'm relationship between risk and expected ret'rn* &n this relationship the expected ret'rn on an asset is a linear f'nction of the asset+s risk with re(ards to other assets and the ret'rn on a riskless asset (e*(* mone,)* Th's.(/n) 0 /i 1 2 Cov (/n- /i) 3here4 .(/n) 0 the expected ret'rn on the nth e)'it, instr'ment /i 0 the ret'rn to a riskless asset Cov (/n- /M) 0 the covariance of /n - the ret'rn of the nth e)'it, instr'ment and /i the rate of ret'rn to the nation+s portfolio of all assets (incl'din( the nth asset)* The new model of capital asset pricin( makes possible a test of the propositions that certain re(ional and scale of iss'e effect biased the allocation of finance passin( thro'(h the British capital market between 18 ! " 1#1$* There are two h,potheses concernin( the bias of British capital flows d'rin( this period* A* British investors and instit'tions were biased in favo'r of overseas assets* B* British investors and instit'tions were biased in favo'r of lar(e borrowers* 5iven the level of competition in the market the implication of these h,potheses is that some factors impin(ed on the British capital market+s perception of risk and its effect on ret'rn* The data from which & draw concl'sions abo't this CAPM 'ses the market for e)'it, shares alone (.delstein- 86)* Tho'(h not incl'din( all t,pes of capital flow- incl'din( nearl, all (overnment debt instr'ments fo'nd in the debent're market- it has broad covera(e in man'fact're and commerce* Method 7sa(e of this pricin( model makes man, ass'mptions* 8irstl, that the British capital market consisted of risk averters who maximised their expected 'tilit, from end of period wealth- and these investors made their optimal investment decisions on the basis of expected ret'rn and standard deviation of ret'rns with the vario's available portfolios* 8'rthermore- it ass'mes that all investors had similar decision hori9ons- and all investors had similar attit'des toward expected ret'rns and the capital market was perfect in the sense that there were man, b',ers and sellers- and that all investors had e)'al access to the same portfolio opport'nities* :ence we can restate the pricin( model as4 .(/n) ; /i 0 2 Cov (/n- /M) .(/n) ; /i is the ret'rn to the nth asset over and above the ret'rn to the riskless asset- or the <risk premi'm+* Cov (/n- /M) is the asset risk variable* /isk on the individ'al asset is hence meas'red b, its contrib'tion to the risk of the national (or market) portfolio- or the asset+s risk with re(ards to other assets* The above relationship between an asset+s risk and expected ret'rn is proportional and the n'merical val'e of the proportion is constant across all assets* 3ith lots of investors operatin( in the capital market in this period- differences in 2 d'e to these two factors were probabl, red'ced to a minim'm d'e to arbitra(e* :owever- if assets of different re(ions or scale were considered to be different commodities- then 2 need not be constant across all assets* The assessment of the capital market in this period contin'es (.delstein 86) b, carr,in( o't statistical tests on the least s)'ares re(ression line of (/= " /c) 0 a1 1 a6> 1 a$?c 1 a Cov(/=-r=m)
Aashish Pattani " /@B&A?@A C@BB.5.
(enerall.18#! " 1#1$* Two proxies for a market rate of ret'rn were tested4 r1m an e)'iproportionate wei(htin( of all ind'strial (ro'pin(s.the textile. less than a million po'nds1 if the =th ind'str.and r 6m.s ne(ative.1 if the =th ind'str.of.chemical.and steel ind'stries were added (?cB)* /es'lts &ndex of market ret'rns 'tilised 18 ! " 8# /1m /6m 18#! " 1#1$ /1m /6m > ?cA ?cB /6 n
.and $ ind'strial (ro'pin(s.* D .confirmin( the direction of bias. was non domestic) ?c 0 (! if the firms of the =th ind'str.onl. was domestic.a$ sho'ld be a ne(ative si(n* The re(ression e)'ation was estimated b. offer little to s'((est an. tr'th in the oft )'oted scale. made iss'es of si(nificantl. railwa. 6 " 1!F in variations in (/ = " /c) G Cov(/=-r=m)* 3ith re(ard to the t test.1*66 1*E8 1*6D
&nsi(* . d'rin( the first decade of the twentieth cent'r.emplo. 'sin( rate of ret'rn observations on $C ind'strial (ro'pin(s18 ! " 188#.+s e)'it.*DE .none of the coefficients pass the t..
. the scale of
Aashish Pattani " /@B&A?@A C@BB.b't wei(hted accordin( to a ro'(h estimate of the market val'e of each ind'str.a different definition of the scale of iss'e (?c) was adopted for each re(ression* 8or the 18 !.the risk premi'm was bid down.drink.and ironcoal.* &n li(ht of the chan(es in the scale of iss'e for ind'strial and commercial companies in the Bondon capital market be(innin( in the mid 188!s.pothesis* The / 6 val'ed indicate that the estimated e)'ation a 1 1 a6> 1 a$?c 1 a explains onl. and social overhead ind'stries were treated as lar(e scale iss'ers (?cA)* &n the 18#! " 1#1$ re(ression. iss'e or re(ional bias h.if the capital market was biased in favo'r of forei(n assets.implied b.tho'(h weak.relative to a 'nit of risk* &f the British capital market was biased in favo'r of lar(e. for a national market rate of ret'rn* > 0 (! if the =th ind'str.scale iss'es.8# test. wealth in the =th ind'str.where /= 0 the compo'nd ann'al interest rate of (rowth of e)'it.in( the same (ro'pin(s as r1m. si(nificant.a 6 sho'ld be ne(ative in si(n* This wo'ld mean that in the act of biddin( for nondomestic assets.invested at time t and sold at time T* /c 0 the compo'nd ann'al rate of (rowth of wealth of a fixed s'm placed in Cons'ls at time t and sold at time T* Cov(/=-r=m) 0 the covariance over time from t to T of / = and r=m a prox.*DE
!*!6 !*!C !*!8 !*1!
$C $C $ $
The res'lts of the re(ression tests for the estimated e)'ation4 (/= " /c) 0 a1 1 a6> 1 a$?c 1 a Cov(/=-r=m) Are fo'nd in the table above* The.*E1 .test criteria at the CF si(nificance level (t val'e of H 6*!D)* Tho'(h none of the coefficients are statisticall.+s iss'es were aro'nd a million po'nds or more) 3ith re(ard to the direction of bias.food.5.their si(ns are s'((estive* The si(n of the scale bias is alwa.
of the re(ression model (explainin( onl.pothesis. no means concl'sive. .delstein* M* . 1!F of the variation).the si(n on the re(ional variable+s coefficient is ne(ative in the 18 ! " 188# re(ression test t'rns positive in the 18#! " 1#1$ test* British investors appear to have switched for a hi(hl.the h.pothesis* &nvestors appear to have had a persistent b't weak tendenc.to a somewhat less 'nstable preference for domestic iss'es.of.does not present a f'll pict're of what was occ'rrin( at the time in British Capital markets* Biblio(raph.18 ! " 1#1$* :owever. their si(ns that biases ma. . are s'((estive b. affected the lon( r'n pricin( of e)'it.b't the simplicit.iss'e h. /* 8lo'd and I*A* McClosk.the res'lts are b.and do not (ive a clear indication of whether biases did or did not exist in 18 ! " 1#1$ British Capital markets* ?o the.?econd edition.5.+ve existed.iss'e biased in the British capital market do not appear to have si(nificantl.pothesi9ed scale. assets.18 ! " 8#.Jol 6Chapter b.domestic iss'es.
.ield lower risk premi'ms per 'nit of risk* Bookin( at the re(ional bias h.(iven the small sample si9e. 'nstable and weak preference for non. to bid 'p the price of lar(e scale iss'es and thereb. The economic history of Britain Since 1700.18#! " 1#1$* Concl'sion 8rom the pricin( model of capital assets proposed in .delstein (1#86).delstein Overseas investment chapter 6
Aashish Pattani " /@B&A?@A C@BB.