does fund size erode mutual fund performance the role of liquidity and organization(3)

Does Fund Size Erode Mutual Fund PerformanceThe Role of Liquidity and Organization

By JOSEPH CHEN HARRISON HONG MING HUANG AND JEFFREY D KUBIK

We investigate the effect of scale on performance in the active money managementindustry We first document that fund returns both before and after fees andexpenses decline with lagged fund size even after accounting for various perfor-mance benchmarks We then explore a number of potential explanations for thisrelationship This association is most pronounced among funds that have to investin small and illiquid stocks suggesting that these adverse scale effects are relatedto liquidity Controlling for its size a fundrsquos return does not deteriorate with the sizeof the family that it belongs to indicating that scale need not be bad for performancedepending on how the fund is organized Finally using data on whether funds aresolo-managed or team-managed and the composition of fund investments weexplore the idea that scale erodes fund performance because of the interaction ofliquidity and organizational diseconomies (JEL G2 G20 G23 L2 L22)

The mutual fund industry plays an increas-ingly important role in the US economy Overthe past two decades mutual funds have beenamong the fastest growing institutions in thiscountry At the end of 1980 they managed lessthan $150 billion but this figure had grown toover $4 trillion by the end of 1997mdasha numberthat exceeds aggregate bank deposits (Robert CPozen 1998) Indeed almost 50 percent ofhouseholds today invest in mutual funds (In-vestment Company Institute 2000) The mostimportant and fastest-growing part of this in-dustry is funds that invest in stocks particularlyactively managed ones The explosion of news-letters magazines and such rating services asMorningstar attest to the fact that investorsspend significant resources in identifying man-agers with stock-picking ability More impor-tant actively managed funds control a sizeable

stake of corporate equity and play a pivotal rolein the determination of stock prices (see egMark Grinblatt et al 1995 Paul Gompers andAndrew Metrick 2001)

In this paper we tackle an issue that is fun-damental to understanding the role of these mu-tual funds in the economymdashthe economies ofscale in the active money management industryNamely how does performance depend on thesize or asset base of the fund A better under-standing of this issue would naturally be usefulfor investors especially in light of the massiveinflows that have increased the mean size offunds in the recent past At the same time theissue of the persistence of fund performancedepends crucially on the scale-ability of fundinvestments (see eg Martin J Gruber 1996Jonathan Berk and Richard C Green 2002)Moreover the nature of the economies of scalein this industry may also have implications forthe agency relationship between managers andinvestors and the optimal compensation con-tract between them (see eg Keith Brown etal 1996 Stan Becker and Greg Vaughn 2001)Therefore understanding the effects of fundsize on fund returns is an important first steptoward addressing such critical issues

While the effect of scale on performance is animportant question it has received little re-search attention to date Some practitioners

Chen Marshall School of Business University ofSouthern California Hoffman Hall 701 Los AngelesCA 90089 (e-mail joechenmarshalluscedu) HongBendheim Center for Finance Princeton University26 Prospect Avenue Princeton NJ 08540 (e-mailhhongPrincetonedu) Huang Stanford Graduate Schoolof Business and Cheong Kong Graduate School of BusinessStanford University Stanford CA 94305 (e-mailmhuangstanfordedu) Kubik Department of EconomicsSyracuse University 426 Eggers Hall Syracuse NY 13244(e-mail jdkubikmaxwellsyredu)

1276

point out that there are advantages to scale suchas more resources for research and lower ex-pense ratios Others believe however that alarge asset base erodes fund performance be-cause of trading costs associated with liquidityor price impact (see eg Andre Perold andRobert S Salomon 1991 Roger Lowenstein1997) Whereas a small fund can easily put allof its money in its best ideas a lack of liquidityforces a large fund to have to invest in itsnot-so-good ideas and take larger positions perstock than is optimal thereby eroding perfor-mance Using a small sample of funds from1974 to 1984 Grinblatt and Sheridan Titman(1989) find mixed evidence that fund returnsdecline with fund size Needless to say there isno consensus on this issue

Using mutual fund data from 1962 to 1999we begin our investigation by using cross-sectional variation to see whether performancedepends on lagged fund size Since funds mayhave different styles we adjust for such heter-ogeneity by utilizing various performancebenchmarks that account for the possibility thatthey load differently on small cap stock valuestock and price momentum strategies More-over fund size may be correlated with suchother fund characteristics as fund age or turn-over and it may be these characteristics that aredriving performance Hence we regress the var-ious adjusted returns on lagged fund size (asmeasured by the log of total net assets undermanagement) and also include in the regres-sions a host of other observable fund character-istics including turnover age expense ratiototal load past-year fund inflows and past-yearreturns A number of studies warn that the re-ported returns of the smallest funds (those withless than $15 million in assets under manage-ment) might be upward biased We excludethese funds from our baseline sample in esti-mating these regressions

The regressions indicate that a fundrsquos perfor-mance is inversely correlated with its laggedassets under management For instance usingmonthly gross returns (before fees and expensesare deducted) a two-standard deviation shockin the log of a fundrsquos total assets under man-agement this month yields anywhere from a 54-to 77-basis-point movement in next monthrsquosfund return depending on the performancebenchmark (or about 65 to 96 basis points an-

nually) The corresponding figures for net fundreturns (after fees and expenses are deducted)are only slightly smaller To put these magni-tudes into some perspective the funds in oursample on average underperform the marketportfolio by about 96 basis points after fees andexpenses From this perspective a 65- to 96-basis-point annual spread in performance is notonly statistically significant but also economi-cally important1

Even after utilizing various performancebenchmarks and controlling for other observ-able fund characteristics there are still a num-ber of potential explanations that might beconsistent with the inverse relationship betweenscale and fund returns To further narrow the setof explanations we proceed to test a directimplication of the hypothesis that fund sizeerodes performance because of trading costsassociated with liquidity and price impact If theldquoliquidity hypothesisrdquo is true then size ought toerode performance much more for funds thathave to invest in small stocks which tend to beilliquid Consistent with this hypothesis we findthat fund size matters much more for the returnsamong such funds identified as ldquosmall caprdquofunds in our database than other funds2 Indeedfor other funds size does not significantly affectperformance This finding strongly indicatesthat liquidity plays an important role in thedocumented diseconomies of scale

We then delve deeper into the liquidity hy-pothesis by observing that liquidity means thatbig funds need to find more stock ideas thansmall funds do but liquidity itself may notcompletely explain why they cannot go aboutdoing this ie why they cannot scale Presum-ably a large fund can afford to hire additionalmanagers to cover more stocks It can therebygenerate additional good ideas so that it can take

1 As we describe below some theories suggest that thesmallest funds may have inferior performance to medium-sized ones because they are being run at a suboptimallysmall scale Because it is difficult to make inferences re-garding the performance of the smallest funds we do notattempt to measure such nonlinearities here

2 Throughout the paper we will sometimes refer to fundswith few assets under management as ldquosmall fundsrdquo andfunds that by virtue of their fund style have to invest insmall stocks as ldquosmall cap fundsrdquo So small cap funds arenot necessarily small funds Indeed most are actually quitelarge in terms of assets under management

1277VOL 94 NO 5 CHEN ET AL DOES FUND SIZE ERODE MUTUAL FUND PERFORMANCE

small positions in lots of stocks as opposed tolarge positions in a few stocks Indeed the vastmajority of stocks with small market capitaliza-tion are untouched by mutual funds (see egHong Lim and Stein 2000 Chen et al 2002)So there is clearly scope for even very largefunds to generate new ideas Put another waywhy canrsquot two small funds (directed by twodifferent managers) merge into one large fundand still have the performance of the large onebe equal to the sum of the two small ones

To see that assets under management neednot be bad for the performance of a fund orga-nization we consider the effect that the size ofthe fund family has on fund performance Manyfunds belong to fund families (eg the famousMagellan fund is part of the Fidelity family offunds) which allows us to measure separatelythe effect of fund size and the size of the rest ofthe family on fund performance Controlling forfund size we find that the assets under manage-ment of the other funds in the family that thefund belongs to actually increase the fundrsquosperformance A two-standard deviation shock tothe size of the other funds in the family leads toabout a 4- to 6-basis-point movement in thefundrsquos performance the following month (orabout 48- to 72-basis-points movement annu-ally) depending on the performance measureused The effect is smaller than that of fund sizeon performance but is nonetheless statisticallyand economically significant As we explain indetail below the most plausible interpretationof this finding is that there are economies asso-ciated with trading commissions and lendingfees at the family level Bigger families likeFidelity are able to get better concessions ontrading commissions and earn higher lendingfees for the stocks held by their funds

These two findingsmdashthat fund performancedeclines with the fundrsquos own size but increaseswith the size of the other funds in the familymdashare both interesting and intuitively appealingFirst in our cross-sectional regressions it isimportant to control for family size in order tofind a sizeable impact of fund size on perfor-mance The reason is that these two variablesare positively correlated and since family sizeis good for performance it is important to con-trol for it to identify the negative effect of fundsize Second the finding on family size alsorules out a number of alternative hypotheses for

our fund size finding For instance it is notlikely that this finding is due to large funds notcaring about returns since large families appar-ently do make sufficient investments to main-tain performance

More important these two findings makeclear that liquidity and scale need not be bad forfund performance per se In most families ma-jor decisions are decentralized in that the fundmanagers make stock picks without substantialcoordination with each other So a family is anorganization that credibly commits to lettingeach of its fund managers run his or her ownassets Moreover being part of a family mayeconomize on certain fixed costs as explainedabove Thus if a large fund is organized like afund family with different managers runningsmall pots of money then scale need not be badper se just as family size does not appear to bebad for family performance

Therefore given that managers care a greatdeal about performance and that scale need notbe bad for performance per se why does itappear that scale erodes fund performance be-cause of liquidity Later in this paper we ex-plore some potential answers to this questionWhereas a small fund can be run by a singlemanager a large fund naturally needs moremanagers and so issues of how the decision-making process is organized become importantWe conjecture that liquidity and scale affectperformance because of certain organizationaldiseconomies We pursue this perspective as ameans to motivate additional analysis involvingfund stock holdings We want to emphasize thatour analysis is exploratory and that a number ofalternative interpretations which we describebelow are possible

There are many types of organizational dis-economies that lead to different predictions onwhy small organizations outperform largeones3 We conjecture that one type known ashierarchy costs (see eg Philippe Aghion andJean Tirole 1997 Jeremy C Stein 2002) maybe especially relevant for mutual funds and mo-tivate our analysis by testing some predictionsfrom Stein (2002) The basic premise is that in

3 See Patrick Bolton and David S Scharfstein (1998) andBengt Holmstrom and John Roberts (1998) for surveys onthe boundaries of the firm that discuss such organizationaldiseconomies

1278 THE AMERICAN ECONOMIC REVIEW DECEMBER 2004

large organizations with hierarchies the processof agents fighting for (and potentially not hav-ing) their ideas implemented will affect agentsrsquoex ante decisions of what ideas they want towork on Stein (2002) argues that in the pres-ence of such hierarchy costs small organiza-tions ought to outperform large ones at tasksthat involve the processing of soft information(ie information that cannot be directly verifiedby anyone other than the agent who producesit) If the information is soft then agents have aharder time convincing others of their ideas andit becomes more difficult to pass this informa-tion up the organization

In the context of mutual funds soft informa-tion most naturally corresponds to research orinvestment ideas related to local stocks (com-panies located near a fund headquarters) sinceanecdotal evidence indicates that investing insuch companies requires that the fund processsoft information eg speaking with CEOs asopposed to simply looking at hard informationlike price-earnings ratios This means that inlarge funds with hierarchies in which managersfight to have their ideas implemented managersmay end up expending too much research efforton quantitative measures of a company (iehard information) so as to convince others toimplement their ideas than they ideally would ifthey controlled their own smaller funds All elseequal large funds may perform worse thansmall ones

Building on the work of Joshua D Coval andTobias J Moskowitz (1999 2001) we find thatconsistent with Stein (2002) small funds espe-cially those investing in small stocks aresignificantly more likely than their larger coun-terparts to invest in local stocks Moreover theydo much better at picking local stocks than largefunds do4

Another implication of Stein (2002) is thatcontrolling for fund size funds that are man-aged by one manager are better at tasks that

involve the processing of soft information thanfunds managed by many managers Consistentwith Stein (2002) we find that solo-managedfunds are significantly more likely than co-managed funds to invest in local stocks andto do better than co-managed funds at pickinglocal stocks Finally we find that controlling forfund size solo-managed funds outperform co-managed funds

Note that such hierarchy costs are not presentat the family level precisely because the familytypically agrees not to reallocate resourcesacross funds Indeed different funds in a familyhave their own boards that deal with such issuesas replacement of managers So the manager incharge of a fund generally does not have toworry about the family taking away the fundrsquosresources and giving them to some other fund inthe family More generally the idea that agentsrsquoincentives are weaker when they do not havecontrol over asset allocation or investment de-cisions is in the work of Sanford J Grossmanand Oliver D Hart (1986) Hart and John Moore(1990) and Hart (1995)

In sum our paper makes a number of contri-butions First we carefully document that per-formance declines with fund size Second weestablish the importance of liquidity in mediat-ing this inverse relationship Third we point outthat the adverse effect of scale on performanceneed not be inevitable because we find thatfamily size actually improves fund perfor-mance Finally we provide some evidence thatthe reason fund size and liquidity do in facterode performance may be due to organizationaldiseconomies related to hierarchy costs It isimportant to note however that our analysisinto the nature of the organizational disecono-mies is exploratory and that there are otherinterpretations which we discuss below

Our paper proceeds as follows In Section Iwe describe the data and in Section II the per-formance benchmarks In Section III we presentour empirical findings We explore alternativeexplanations in Section IV and conclude inSection V

I Data

Our primary data on mutual funds come fromthe Center for Research in Security Prices(CRSP) Mutual Fund Database which spans the

4 Steinrsquos analysis also suggests that large organizationsneed not underperform small ones when it comes to pro-cessing hard information In the context of the mutual fundindustry only passive index funds like Vanguard are likelyto rely solely on hard information Most active mutual fundsrely to a significant degree on soft information Interest-ingly anecdotal evidence indicates that scale is not as big anissue for passive index funds as it is for active mutual funds


years 1962 to 1999 Following many prior stud-ies we restrict our analysis to diversified USequity mutual funds by excluding from our sam-ple bond international and specialized sectorfunds5 For a fund to be in our sample it mustreport information on assets under managementand monthly returns We require also that ithave at least one year of reported returns Thisadditional restriction is imposed because we needto form benchmark portfolios based on past fundperformance6 Finally a mutual fund may enterthe database multiple times in the same month if ithas different share classes We clean the data byeliminating such redundant observations

Table 1 reports summary statistics for oursample In panel (A) we report the means andstandard deviations for the variables of interestfor each fund size quintile for all funds and forfunds in fund size quintiles two (next to small-est) to five (largest) Edwin J Elton et al (2001)warn that one has to be careful in making in-ferences regarding the performance of fundsthat have less than $15 million in total net assetsunder management They point out that there isa systematic upward bias in the reported returnsamong these observations This bias is poten-tially problematic for our analysis since we areinterested in the relationship between scale andperformance As we will see shortly this cri-tique applies only to observations in fund sizequintile one (smallest) since the funds in theother quintiles typically have greater than $15million under management Therefore we focusour analysis on the subsample of funds in fundsize quintiles two through five It turns out thatour results are robust whether or not we includethe smallest funds in our analysis

We utilize 3439 distinct funds and a total

27431 fund years in our analysis7 In eachmonth our sample includes on average 741funds They have average total net assets (TNA)of $2825 million with a standard deviation of$9258 million The interesting thing to notefrom the standard deviation figure is that there isa substantial spread in TNA Indeed this be-comes transparent when we disaggregate thesestatistics by fund size quintiles Those in thesmallest quintile have an average TNA of onlyabout $47 million whereas the ones in the topquintile have an average TNA of over $11 bil-lion The funds in fund size quintiles twothrough five have a slightly higher mean of$3523 million with a standard deviation of over$1 billion For the usual reasons related toscaling the proxy of fund size that we will usein our analysis is the log of a fundrsquos TNA(LOGTNA) The statistics for this variable arereported in the row below that of TNA Anothervariable of interest is LOGFAMSIZE which isthe log of one plus the cumulative TNA of theother funds in the fundrsquos family (ie the TNAof a fundrsquos family excluding its own TNA)

In addition the database reports a host ofother fund characteristics that we utilize in ouranalysis The first is fund turnover (TURN-OVER) defined as the minimum of purchasesand sales over average TNA for the calendaryear The average fund turnover is 542 percentper year The average fund age (AGE) is about157 years The funds in our sample have ex-pense ratios as a fraction of year-end TNA (EX-PRATIO) that average about 97 basis points peryear They charge a total load (TOTLOAD) ofabout 436 percent (as a percentage of newinvestments) on average FLOW in month t isdefined as the fundrsquos TNA in month t minus theproduct of the fundrsquos TNA at month t 12 withthe net fund return between months t 12 andt all divided by the fundrsquos TNA at month t 12The funds in the sample have an average fundflow of 247 percent per year These summarystatistics are similar to those obtained for thesubsample of funds in fund size quintiles twothrough five

5 More specifically we select mutual funds in the CRSPMutual Fund Database that have reported one of the fol-lowing investment objectives at any point We first selectmutual funds with the Investment Company Data Inc(ICDI) mutual fund objective of ldquoaggressive growthrdquoldquogrowth and incomerdquo or ldquolong-term growthrdquo We then addin mutual funds with the Strategic Insight mutual fundobjective of ldquoaggressive growthrdquo ldquoflexiblerdquo ldquogrowth andincomerdquo ldquogrowthrdquo ldquoincome-growthrdquo or ldquosmall companygrowthrdquo Finally we select mutual funds with the Wiesen-berger mutual fund objective code of ldquoGrdquo ldquoG-Irdquo ldquoG-I-SrdquoldquoG-Srdquo ldquoGCIrdquo ldquoI-Grdquo ldquoI-S-Grdquo ldquoMCGrdquo or ldquoSCGrdquo

6 We have also replicated our analysis without this re-striction The only difference is that the sample includesmore small funds but the results are unchanged

7 At the end of 1993 we have about 1508 distinct fundsin our sample very close to the number reported by previ-ous studies that have used this database Moreover thesummary statistics below are similar to those reported inthese other studies as well


TABLE 1mdashSUMMARY STATISTICS

Panel A Time-series averages of cross-sectional averages and standard deviations

Mutual fund size quintile

All funds Quintiles 2ndash51 2 3 4 5

Number of funds 1482 1478 1478 1478 1473 7414 5906TNA 47 222 606 1654 11647 2825 3523

($ million) [32] [72] [166] [548] [17971] [9258] [10227]LOGTNA 109 294 396 494 645 387 457

($ million) [101] [034] [027] [033] [075] [192] [138]LOGFAMSIZE 370 452 529 591 700 528 568

($ million) [298] [289] [272] [256] [216] [292] [275]TURNOVER 4207 5568 5909 6120 5217 5417 5707

( per year) [8303] [7400] [6860] [6428] [5456] [7184] [6721]AGE 817 1190 1482 1843 2516 1567 1757

(years) [854] [1073] [1285] [1438] [1506] [1396] [1433]EXPRATIO 129 108 094 085 068 097 089

( per year) [111] [059] [046] [037] [031] [068] [048]TOTLOAD 341 419 432 457 528 436 459

() [332] [332] [334] [339] [288] [336] [332]FLOW 3079 3066 2697 2127 1354 2467 2313

( per year) [11355] [11336] [10166] [8408] [5904] [10264] [9660]

Panel B Time-series averages of (monthly) correlations between fund characteristics (using all funds)

TNA LOGTNA LOGFAMSIZE TURNOVER AGE EXPRATIO TOTLOAD FLOW

TNA 100 056 024 005 027 019 010 003LOGTNA 100 040 006 044 031 019 007LOGFAMSIZE 100 008 008 019 025 001TURNOVER 100 001 017 005 001AGE 100 013 019 018EXPRATIO 100 005 008TOTLOAD 100 004FLOW 100

Panel C Time-series averages of (monthly) correlations between fund characteristics (excluding smallest 20 percent of funds)



Panel D Time-series averages of (monthly) cross-sectional averages of market-adjusted fund returns



FUNDRET 009 002 003 006 006 001 002(Gross) [304] [264] [261] [246] [200] [262] [248]

FUNDRET 002 007 005 013 012 008 009(Net) [304] [264] [261] [246] [200] [262] [248]

Notes This table reports summary statistics for the funds in our sample ldquoNumber of fundsrdquo is the number of mutual funds that meet our selectioncriteria for being an active mutual fund in each month TNA is the total net assets under management in millions of dollars LOGTNA is the logarithmof TNA LOGFAMSIZE is the logarithm of one plus the assets under management of the other funds in the family that the fund belongs to (excludingthe asset base of the fund itself) TURNOVER is fund turnover defined as the minimum of aggregate purchases and sales of securities divided by theaverage TNA over the calendar year AGE is the number of years since the establishment of the fund EXPRATIO is the total annual management feesand expenses divided by year-end TNA TOTLOAD is the total front-end deferred and rear-end charges as a percentage of new investments FLOWis the percentage new fund flow into the mutual fund over the past year TNA LOGFAMSIZE and FLOW are reported monthly All other fundcharacteristics are reported once a year FUNDRET is the monthly market-adjusted fund return These returns are calculated before (gross) and after(net) deducting fees and expenses Panel (A) reports the time-series averages of monthly cross-sectional averages and monthly cross-sectional standarddeviations (shown in brackets) of fund characteristics Panels (B) and (C) report the time-series averages of the cross-sectional correlations betweenfund characteristics Panel (D) reports the time-series averages of monthly cross-sectional averages of market-adjusted fund returns In panels (A) and(B) fund size quintile 1 (5) has the smallest (largest) funds The sample is from January 1963 to December 1999


Panel (B) of Table 1 reports the time-seriesaverages of the cross-sectional correlations be-tween the various fund characteristics A num-ber of patterns emerge First LOGTNA isstrongly correlated with LOGFAMSIZE (040)Second EXPRATIO varies inversely with LOG-TNA (031) while TOTLOAD and AGE varypositively with LOGTNA (019 and 044 re-spectively) Panel (C) reports the analogousnumbers for the funds in fund size quintiles twothrough five The results are similar to those inpanel (B) It is apparent from panels (B) and (C)that we need to control for these fund charac-teristics in estimating the cross-sectional rela-tionship between fund size and performance

Finally we report in panel (D) the means andstandard deviations for the monthly fund re-turns FUNDRET where we measure these re-turns in a couple of different ways We firstreport summary statistics for gross fund returnsadjusted by the return of the market portfolio(simple market-adjusted returns) Monthlygross fund returns are calculated by adding backthe expenses to net fund returns by taking theyear-end expense ratio dividing it by 12 andadding it to the monthly returns during the yearFor the whole sample the average monthly per-formance is 1 basis point with a standard devi-ation of 262 percent The funds in fund sizequintiles two through five do slightly worsewith a mean of 2 basis points and a standarddeviation of 248 percent We also report thesesummary statistics using net fund returns Thefunds in our sample underperform the market by8 basis points per month or 96 basis points peryear after fees and expenses are deducted

These figures are almost identical to thosedocumented in other studies These studies findthat fund managers do have the ability to beat orstay even with the market before managementfees (see eg Grinblatt and Titman 1989Grinblatt et al 1995 Kent Daniel et al 1997)However mutual fund investors are apparentlywilling to pay a lot in fees for limited stock-picking ability which results in their risk-adjusted fund returns being significantlynegative (see eg Michael C Jensen 1968Burton G Malkiel 1995 Gruber 1996)

Moreover notice that smaller funds appear tooutperform their larger counterparts For in-stance funds in quintile two have an averagemonthly gross return of 2 basis points while

funds in quintile five underperform the marketby 6 basis points The difference of 8 basispoints per month or 96 basis points a year is aneconomically interesting number Net fund re-turns also appear to be negatively correlatedwith fund size though the spread is somewhatsmaller than using gross returns We do notwant to overinterpret these results since we havenot controlled for heterogeneity in fund stylesnor calculated any type of statistical signifi-cance in this table

In addition to the CRSP Mutual Fund Data-base we also utilize the CDA Spectrum Data-base to analyze the effect of fund size on thecomposition of fund stock holdings and theperformance of these holdings The reason weneed to augment our analysis with this databaseis that the CRSP Mutual Fund Database doesnot contain information on fund positions inindividual stocks The CDA Spectrum Databasereports a fundrsquos stock positions on a quarterlybasis but it is not available until the early 1980sand does not report a fundrsquos cash positionsRuss Wermers (2000) compared the funds inthese two databases and found that the activefunds represented in the two databases are com-parable So while the CDA Spectrum Databaseis less effective than the CRSP Mutual FundDatabase in measuring performance it is ade-quate for analyzing the effects of fund size onstock positions We will provide a more detaileddiscussion of this database in Section III

II Methodology

Our empirical strategy utilizes cross-sectionalvariation to see how fund performance varieswith lagged fund size We could have adopted afixed-effects approach by looking at whetherchanges in a fundrsquos performance are related tochanges in its size Such an approach is subjecthowever to a regression-to-the-mean bias Afund with a year or two of lucky performancewill experience an increase in fund size Butperformance will regress to the mean leading toa spurious conclusion that an increase in fundsize is associated with a decrease in fund re-turns Measuring the effect of fund size onperformance using cross-sectional regressionsis less subject to such bias Indeed it may evenbe conservative given our goal since largerfunds are likely to be better funds or they would


not have become large in the first place We arelikely to be biased toward finding any disecono-mies of scale using cross-sectional variation

There are two major worries that arise how-ever when using cross-sectional variation Thefirst is that funds of different sizes may be indifferent styles For instance small funds mightbe more likely than large funds to pursue smallstock value stock and price momentum strate-gies which have been documented to generateabnormal returns While it is not clear that onenecessarily wants to adjust for such heterogene-ity it would be more interesting if we found thatpast fund size influences future performanceeven after accounting for variations in fundstyles The second worry is that fund size mightbe correlated with other fund characteristicssuch as fund age or turnover and it may bethese characteristics that are driving perfor-mance For instance fund size may be measur-ing whether a fund is active or passive (whichmay be captured by fund turnover) While wehave tried our best to rule out passive funds inour sample construction it is possible that somefunds may just be indexers And if it turns outthat indexers happen to be large funds becausemore investors put their money in such fundsthen size may be picking up differences in thedegree of activity among funds

A Fund Performance Benchmarks

A very conservative way to deal with the firstworry about heterogeneity in fund styles is toadjust for fund performance by various bench-marks In this paper we consider in addition tosimple market-adjusted returns returns adjustedby the Capital Asset Pricing Model (CAPM) ofWilliam F Sharpe (1964) We also considerreturns adjusted using the Eugene F Famaand Kenneth R French (1993) three-factormodel and this model augmented with themomentum factor of Narasimhan Jegadeesh andTitman (1993) which has been shown invarious contexts to provide explanatory powerfor the observed cross-sectional variation infund performance (see eg Mark M Carhart1997)

Panel (A) of Table 2 reports the summarystatistics for the various portfolios that make upour performance benchmarks Among theseare the returns on the CRSP value-weighted

stock index net of the one-month Treasuryrate (VWRF) the returns to the Fama andFrench (1993) SMB (small stocks minus largestocks) and HML (high book-to-market stocksminus low book-to-market stocks) portfoliosand the returns-to-price momentum portfolioMOM12 (a portfolio that is long stocks that arepast-12-month winners and short stocks that arepast-12-month losers and hold for one month)The summary statistics for these portfolio re-turns are similar to those reported in other mu-tual fund studies

Since we are interested in the relationshipbetween fund size and performance we sortmutual funds at the beginning of each monthbased on the quintile rankings of their previous-month TNA8 We then track these five portfoliosfor one month and use the entire time series oftheir monthly net returns to calculate the load-ings to the various factors (VWRF SMB HMLMOM12) for each of these five portfolios Foreach month each mutual fund inherits the load-ings of the one of these five portfolios that itbelongs to In other words if a mutual fundstays in the same-size quintile throughout itslife its loadings remain the same But if itmoves from one size quintile to another duringa certain month it inherits a new set of load-ings with which we adjust its next monthrsquosperformance

Panel (B) reports the loadings of the fivefund-size (TNA) sorted mutual fund portfoliosusing the CAPM

(1)

Rit i i VWRFt it t 1 T

where Rit is the (net fund) return on one of ourfive fund-size-sorted mutual fund portfolios inmonth t in excess of the one-month T-bill re-turn i is the excess return of that portfolio iis the loading on the market portfolio and itstands for a generic error term that is uncorre-lated with all other independent variables As

8 We also sort mutual funds by their past-12-month re-turns to form benchmark portfolios Our results are un-changed when using these benchmark portfolios We omitthese results for brevity


other papers have found the average mutualfund has a beta of around 091 reflecting thefact that mutual funds hold some cash or bondsin their portfolios Notice that there is only aslight variation in the market beta (irsquos) fromthe smallest to the largest fund size portfoliothe smallest portfolio has a somewhat smallerbeta but not by much

Panel (C) reports the loadings for two addi-tional performance models the Fama-Frenchthree-factor model and this three-factor modelaugmented by a momentum factor

(2) Rit i i1 VWRFt i2 SMBt

i3 HMLt it t 1 T


i3 HMLt i4 MOM12t it

t 1 T

where Rit is the (net fund) return on one of ourfive size-sorted mutual fund portfolios in montht in excess of the one-month T-bill return i isthe excess return irsquos are loadings on the var-ious portfolios and it stands for a generic errorterm that is uncorrelated with all other indepen-dent variables We see that small funds tend tohave higher loadings on SMB and HML butlarge funds tend to load a bit more on momen-tum For instance the loading on SMB for the

TABLE 2mdashSUMMARY STATISTICS FOR PERFORMANCE BENCHMARKS

Panel A Summary statistics of the factors

FactorMeanreturn

SD ofreturn

Cross-correlations

VWRF SMB HML MOM12

VWRF 058 437 100 032 039 002SMB 017 290 100 016 030HML 034 263 100 015MOM12 096 388 100

Panel B Loadings calculated using the CAPM

Portfolio

CAPM

Alpha VWRF

1 (small) 004 0872 002 0913 001 0934 009 0925 (large) 007 091

Panel C Loadings calculated using the 3-Factor model and the 4-Factor model

Portfolio

3-Factor model

HML

4-Factor model

MOM12Alpha VWRF SMB Alpha VWRF SMB HML

1 (small) 001 082 029 003 002 082 030 004 0032 003 084 027 001 009 085 030 000 0053 001 087 022 005 006 087 025 003 0064 006 087 018 005 013 087 020 004 0065 (large) 005 088 008 006 010 088 010 005 005

Notes This table reports the loadings of the five (equal-weighted) TNA-sorted fund portfolios on various factors Panel (A)reports the summary statistics for the factors VWRF is the return on the CRSP value-weighted stock index in excess of theone-month Treasury rate SMB is the return on a portfolio of small stocks minus large stocks HML is the return on a portfoliolong high book-to-market stocks and short low book-to-market stocks MOM12 is the return on a portfolio long stocks thatare past-12-month winners and short those that are past-12-month losers Panel (B) reports the loadings calculated using theCAPM Panel (C) reports the loadings calculated using the Fama-French (1993) 3-Factor model and this model augmentedwith the momentum factor (4-Factor model) The sample period is from January 1963 to December 1999


three-factor model for funds in quintile one is029 while the corresponding loading for fundsin quintile five is 008 And whereas large fundsload negatively on HML (006 for the largestfunds) the smallest funds load positively onHML (003) (Eric G Falkenstein [1996] alsofinds some evidence that larger funds tend toplay large and glamour stocks by looking atfund holdings)

We have also redone all of our analysis bycalculating these loadings using gross fund re-turns instead of net fund returns As the resultsare very similar to using net fund returns forbrevity we will use the loadings summarized inTable 2 to adjust fund performance below(whether it be gross or net returns) Using theentire time series of a particular fund (we re-quire at least 36 months of data) we also cal-culate the loadings separately for each mutualfund using equations (1) to (3) This techniqueis not as good in the sense that we have a muchmore selective requirement on selection and theestimated loadings tend to be very noisy In anycase our results are unchanged so we omitthese results for brevity

B Regression Specifications

To deal with the second concern related to thecorrelation of fund size with other fund charac-teristics we analyze the effect of past fund sizeon performance in the regression frameworkproposed by Fama and James D MacBeth(1973) where we can control for the effects ofother fund characteristics on performance Spe-cifically the regression specification that weutilize is

(4) FUNDRETit LOGTNAit 1

Xit 1 it i 1 N

where FUNDRETit is the return (either gross ornet) of fund i in month t adjusted by variousperformance benchmarks is a constant LOG-TNAit1 is the measure of fund size and Xit1is a set of control variables (in month t 1) thatincludes LOGFAMSIZEit1 TURNOVERit1AGEit1 EXPRATIOit1 TOTLOADit1 andFLOWit1 In addition we include in the right-hand side LAGFUNDRETit1 which is thepast-year return of the fund Here it again

stands for a generic error term that is uncorre-lated with all other independent variables Thecoefficient of interest is which captures therelationship between fund size and fund perfor-mance controlling for other fund characteris-tics is the vector of loadings on the controlvariables We then take the estimates from thesemonthly regressions and follow Fama and Mac-Beth (1973) in taking their time series meansand standard deviations to form our overall es-timates of the effects of fund characteristics onperformance

We will also utilize an additional regressionspecification given by the following

(5) FUNDRETit 1 LOGTNAit 1

2 IndStyle 3LOGTNAit 1 IndStyle

Xit 1 it i 1 N

where the dummy indicator IndStyle (thatequals one if a fund belongs to a certain stylecategory and zero otherwise) and the remainingvariables are the same as in equation (3) Thecoefficient of interest is 3 which measures thedifferential effect of fund size on returns acrossdifferent fund styles It is important to note thatwe do not attempt to measure whether the rela-tionship between fund performance and fundsize may be nonlinear While some theoriesmight suggest that very small funds may haveinferior performance to medium-sized ones be-cause they are being operated at a suboptimallysmall scale we are unable to get at this issuebecause inference regarding the performance ofthe smallest funds is problematic for the reasonsarticulated in Section II

III Results

A Relationship between Fund Size andPerformance

In Table 3 we report the estimation resultsfor the baseline regression specification given inequation (4) We begin by reporting the resultsfor gross fund returns The sample consists offunds from fund size quintiles two through fiveNotice that the coefficient in front of LOGTNAis negative and statistically significant acrossthe four performance measures The coefficients


obtained using either market-adjusted orCAPM-adjusted returns are around 0028with t-statistics of around three Since one stan-dard deviation of LOGTNA is 138 a two-stan-dard deviation shock to fund size means thatperformance changes by 0028 times 28 or 8basis points per month (96 basis points peryear) For the other two performance bench-marks the 3-factor and 4-factor adjusted re-turns the coefficients are slightly smaller at002 but both are still statistically significantwith t-statistics of between 21 and 25 Forthese coefficients a two-standard deviationshock to fund size means that performancechanges by around 70 basis points annually

To put these magnitudes into some perspec-tive observe that a standard deviation of mutualfund returns is around 10 percent annually withslight variations around this figure dependingon the performance measure Hence a two-

standard deviation shock in fund size yields amovement in next yearrsquos fund return that isapproximately 10 percent of the annual volatil-ity of mutual funds (96 basis points divided by10 percent) Another way to think about thesemagnitudes is that the typical fund has a grossfund performance net of the market return thatis basically near zero As a result a spread infund performance of anywhere from 70 to 96basis points a year is quite economicallysignificant

Table 3 also reveals a number of other inter-esting findings The only other variables besidefund size that are statistically significant areLOGFAMSIZE and LAGFUNDRET Interest-ingly LOGFAMSIZE predicts better fund per-formance We will have much more to sayabout the coefficient in front of LOGFAMSIZElater It should be emphasized at this point thatit is important to control for family size in order

TABLE 3mdashREGRESSION OF FUND PERFORMANCE ON LAGGED FUND SIZE

Gross fund returns Net fund returns

Market-Adj Beta-Adj 3-Factor 4-Factor Market-Adj Beta-Adj 3-Factor 4-Factor

INTERCEPT 0056 0087 0080 0019 0026 0056 0049 0011(096) (163) (128) (030) (044) (105) (079) (018)

LOGTNAit1 0028 0028 0023 0020 0025 0025 0020 0018(302) (304) (242) (215) (275) (276) (216) (189)

LOGFAMSIZEit1 0007 0007 0007 0007 0007 0007 0007 0007(226) (227) (222) (221) (233) (234) (229) (228)

TURNOVERit1 0000 0000 0000 0000 0000 0000 0000 0000(085) (086) (085) (083) (077) (078) (077) (075)

AGEit1 0001 0001 0001 0001 0000 0000 0001 0001(062) (062) (064) (063) (052) (053) (055) (054)

EXPRATIOit1 0004 0004 0007 0007 0039 0038 0041 0041(011) (009) (018) (018) (097) (095) (104) (105)

TOTLOADit1 0003 0003 0003 0003 0003 0003 0003 0003(126) (125) (126) (129) (121) (120) (121) (125)

FLOWit1 0000 0000 0000 0000 0000 0000 0000 0000(050) (050) (051) (049) (047) (047) (048) (046)

LAGFUNDRETit1 0029 0028 0028 0029 0029 0029 0029 0029(600) (598) (600) (599) (603) (601) (603) (602)

No of months 444 444 444 444 444 444 444 444

Notes This table shows the Fama-MacBeth (1973) estimates of monthly fund returns regressed on fund characteristics laggedone month The sample includes only funds that fall within fund size quintiles two to five Fund returns are calculated before(gross) and after (net) deducting fees and expenses These returns are adjusted using the market model the CAPM theFama-French (1993) 3-Factor model and the 4-Factor model The dependent variable is FUNDRET LOGTNA is the naturallogarithm of TNA LOGFAMSIZE is the natural logarithm of one plus the size of the family that the fund belongs toTURNOVER is fund turnover and AGE is the number of years since the organization of the mutual fund EXPRATIO is thetotal annual management fees and expenses divided by TNA TOTLOAD is the total front-end deferred and rear-end chargesas a percentage of new investments FLOW is the percentage new fund flow into the mutual fund over the past one yearLAGFUNDRET is the cumulative (buy-hold) fund return over the past 12 months The sample is from January 1963 toDecember 1999 The t-statistics are adjusted for serial correlation using Newey-West (1987) lags of order three and are shownin parentheses


to find a sizeable impact of fund size on perfor-mance This is because fund and family size arepositively correlated and since family size isgood for performance it is important to controlfor it to identify the negative effect of fund sizeThis is a major reason why other studies thathave had fund size as a control variable inexplaining fund returns do not find any signifi-cant effect As a result these studies mentionfund size only in passing and do not provide anyanalysis whatsoever

The fact that the coefficient in front ofLAGFUNDRET is significant suggests thatthere is some persistence in fund returns As forthe rest of the variables some come in withexpected signs though none is statistically sig-nificant The coefficient in front of EXPRATIOis negative consistent with industry observa-tions that larger funds have lower expense ra-tios The coefficients in front of TOTLOAD andTURNOVER are positive as these two variablesare thought to be proxies for whether a fund isactive or passive Fund flow has a negligibleability to predict fund returns and the age of thefund comes in with a negative sign but is sta-tistically insignificant

We next report the results of the baselineregression using net fund returns The coeffi-cient in front of LOGTNA is still negative andstatistically significant across all performancebenchmarks Indeed the coefficient in front ofLOGTNA is only slightly smaller using net fundreturns than using gross fund returns The ob-servations regarding the economic significanceof fund size made earlier continue to hold Ifanything they are even more relevant in thiscontext since the typical fund tends to under-perform the market by about 96 basis pointsannually The coefficients in front of the othervariables have similar signs to those obtainedusing gross fund returns Importantly keep inmind that the coefficient in front of LOGFAM-SIZE is just as statistically and economicallysignificant using net fund returns as gross fundreturns

In Table 4 we present various permutationsinvolving the regression specification in equa-tion (4) to see if the results in Table 3 are robustIn panel (A) we present the results using all thefunds in our sample including those in thesmallest fund size quintile As we mentionedearlier the performance of the funds in the

bottom fund size quintile is biased upward sowe should not draw too much from this analysisother than that our results are unchanged byincluding them in the sample For brevity wereport only the coefficients in front of LOGTNAand LOGFAMSIZE Using gross fund returnsthe coefficient in front of LOGTNA ranges from0019 to 0026 depending on the perfor-mance measure For net fund returns it rangesfrom 0015 to 0022 All the coefficients arestatistically significant at the 5-percent levelwith the exception of the coefficient obtainedusing 3-factor adjusted net fund returns Thecoefficient in this instance is significant only atthe 10-percent level of significance The mag-nitudes are somewhat smaller using the fullsample than the sample that excludes the small-est quintile This difference however is notlarge Moreover the coefficients in front ofLOGFAMSIZE are similar in magnitude tothose obtained in Table 3 As such we concludethat our key findings in Table 3 are robust toincluding all funds in the sample

In panel (B) we attempt to predict a fundrsquoscumulative return next year rather than its returnnext month Not surprisingly we find similarresults to those in Table 3 The coefficient infront of LOGTNA is negative and statisticallysignificant across all performance benchmarksIndeed the economic magnitudes implied bythese estimates are similar to those obtained inTable 3 These statements apply equally toLOGFAMSIZE

In panels (C) and (D) we split our bench-mark sample in half to see whether our esti-mates on LOGTNA and LOGFAMSIZE dependon particular subperiods 1963 to 1980 and 1981to 1999 It appears that LOGTNA has a strongnegative effect on performance regardless of thesubperiods since the economic magnitudes arevery similar to those obtained in Table 3 Wewould not be surprised if the coefficients werenot statistically significant since we havesmaller sample sizes in panels (C) and (D) Buteven with only half the sample size LOGTNAcomes in significantly for a number of the per-formance measures In contrast it appears thatthe effect of LOGFAMSIZE on performance ismuch more pronounced in the latter half of thesample

The analyses in Tables 3 and 4 strongly in-dicate that fund size is negatively related to


future fund performance Moreover we are ableto rule out that this relationship is driven bydifferences in fund styles or mechanical corre-lations of fund size with other observable fundcharacteristics However there still remain a

number of potential explanations for thisrelationship

Three explanations come to mind First thelagged fund size and performance relationshipis due to transaction costs associated with li-

TABLE 4mdashREGRESSION OF FUND RETURNS ON LAGGED FUND SIZE ROBUSTNESS CHECKS

Panel A Sample includes all funds



LOGTNAit1 0023 0026 0019 0021 0020 0022 0015 0017(259) (294) (222) (248) (216) (249) (176) (202)

LOGFAMSIZEit1 0006 0006 0006 0006 0006 0006 0006 0006(199) (203) (199) (205) (210) (215) (211) (217)

No of months 444 444 444 444 444 444 444 444

Panel B Dependent variable is 12-month (non-overlapping) fund returns



LOGTNAit1 0436 0440 0345 0312 0402 0406 0312 0280(326) (327) (240) (219) (307) (307) (220) (199)

LOGFAMSIZEit1 0088 0089 0088 0088 0090 0091 0090 0090(205) (206) (205) (207) (210) (212) (211) (212)

No of years 37 37 37 37 37 37 37 37

Panel C Sample period is from 1963 to 1980



LOGTNAit1 0035 0034 0021 0019 0032 0032 0019 0016(241) (242) (151) (133) (222) (223) (132) (115)

LOGFAMSIZEit1 0002 0002 0002 0002 0002 0002 0002 0002(049) (050) (045) (044) (056) (058) (053) (052)

No of months 216 216 216 216 216 216 216 216

Panel D Sample period is from 1981 to 1999



LOGTNAit1 0021 0022 0024 0022 0019 0019 0022 0019(184) (186) (193) (171) (164) (165) (174) (153)

LOGFAMSIZEit1 0012 0012 0012 0012 0012 0012 0012 0012(253) (253) (250) (250) (255) (255) (252) (252)

No of months 228 228 228 228 228 228 228 228

Notes This table presents robustness checks of the regression specification in Table 3 In panel (A) the sample includes allfunds In panel (B) the dependent variable is 12-month fund returns and the regressions are non-overlapping In panel (C)the sample consists only of observations from 1963 to 1980 In panel (D) the sample consists only of observations from 1981to 1999 LOGTNA is the natural logarithm of TNA LOGFAMSIZE is the natural logarithm of one plus the size of the familythat the fund belongs to Estimates of the intercept and other independent variables are omitted for brevity The otherindependent variables include TURNOVER AGE EXPRATIO TOTLOAD FLOW and LAGFUNDRET The t-statistics areadjusted for serial correlation using Newey-West (1987) lags of order three and are shown in parentheses


quidity or price impact We call this the ldquoliquid-ity hypothesisrdquo Second perhaps investors inlarge funds are less discriminating about returnsthan investors in small funds One reason whythis might be the case is that such large funds asMagellan are better at marketing and are ableto attract investors through advertising Incontrast small funds without such marketingoperations may need to rely more on perfor-mance to attract and maintain investors Wecall this the ldquoclientele hypothesisrdquo Thirdfund size is inversely related to performancebecause of fund incentives to lock in assetsunder management after a long string of goodpast performances9 When a fund is small andhas a limited reputation the manager goesabout the business of stock picking But as thefund gets large because of good past perfor-mance the manager may for various reasonslock in his fund size by being passive (orbeing a ldquocloset indexerrdquo as practitioners putit) We call this the ldquoagency-risk-taking hy-pothesisrdquo The general theme behind the sec-ond and third hypotheses is that after fundsreach a certain size they no longer care aboutmaximizing returns

B The Role of Liquidity Fund Size FundStyles and the Number of Stocks in a Fundrsquos

Portfolio

In order to narrow the list of potential expla-nations we design a test of the liquidity hypoth-esis To the extent that liquidity is driving ourfindings above we would expect to see thatfund size matters much more for performanceamong funds that have to invest in small stocks(ie stocks with small market capitalization)than funds that get to invest in large stocks The

reason is that small stocks are notoriously illiq-uid As a result funds that have to invest insmall stocks are more likely to need new stockideas with asset base growth whereas largefunds can simply increase their existing posi-tions without being hurt too much by priceimpact

Importantly this test of the liquidity hypoth-esis also allows us to discriminate between theother two hypotheses First existing researchfinds that there is little variation in incentivesbetween ldquosmall caprdquo funds (ie funds that haveto invest in small stocks) and other funds (seeeg Andres Almazan et al 2001) Hence thisprediction ought to help us discriminate be-tween our hypothesis and the alternativeldquoagency-risk-takingrdquo story involving fund in-centives Moreover since funds that have toinvest in small stocks tend to do better thanother funds it is not likely that our results aredue to the clientele of these funds being moreirrational than those investing in other fundsThis allows us to distinguish the liquidity storyfrom the clientele story

In the CRSP Mutual Fund Database we arefortunate that each fund self-reports its styleand so we looked for style descriptions contain-ing the words ldquosmall caprdquo It turns out that onestyle ldquoSmall Cap Growthrdquo fits this criterionFunds in this category are likely to have toinvest in small stocks by virtue of their styledesignation Thus we identify a fund in oursample as either Small Cap Growth if it has everreported itself as such or ldquoNot Small CapGrowthrdquo (Funds rarely change their self-reported style) Unfortunately funds with thisdesignation are not prevalent until the early1980s Therefore throughout the analysis inthis section we limit our sample to the 1981ndash1999 period During this period there are onaverage 165 such funds each year The corre-sponding number for the overall sample duringthis period is about 1000 Therefore Small CapGrowth represents a small but healthy slice ofthe overall population Also the average TNA ofthese funds is $2129 million with a standarddeviation of $5667 million The average TNAof a fund in the overall sample during this latterperiod is $4315 million with a standard devia-tion of $158 billion Thus Small Cap Growthfunds are smaller than the typical fund But theyare still quite big and there is a healthy fund size

9 More generally it may be that after many years of goodperformance bad performance follows for whatever reasonWe are offering here a plausible economic mechanism forwhy this might come about The ex ante plausibility of thisalternative story is however somewhat mixed On the onehand the burgeoning empirical literature on career concernssuggests that fund managers ought to be bolder with pastsuccess (see eg Judith A Chevalier and Glenn D Ellison1999 and Hong Kubik and Solomon 2000) On the otherhand the fee structure means that funds may want to lock inassets under management because investors are typicallyslow to pull their money out of funds (Brown et al 1996Chevalier and Ellison 1997)


distribution among them so that we can mea-sure the effect of fund size on performance

Table 5 panel (A) reports what happens tothe results in Table 3 when we augment theregression specifications by including adummy indicator Indnot SCG (which equalsone if a fund is not Small Cap Growth andzero otherwise) and an additional interactionterm involving LOGTNA and Indnot SCG asin equation (5) We first report the results forgross fund returns The coefficient in front ofLOGTNA is about 006 (across the fourperformance benchmarks) Importantly thecoefficient in front of the interaction term ispositive and statistically significant (about004 across the four performance bench-marks) This is the sign predicted by the li-quidity hypothesis since it says that for NotSmall Cap Growth funds there is a smallereffect of fund size on performance The effectis economically interesting as well Since thetwo coefficients 006 and 004 are similarin magnitude a sizeable fraction of the effectof fund size on performance comes fromsmall cap funds The results using net fundreturns reported are similar

In panel (A) of Table 5 we compared SmallCap Growth funds to all other funds In panel(B) of Table 5 we delve a bit deeper into theldquoliquidity hypothesisrdquo by looking at how fundperformance varies with fund size dependingon whether a fund is a Large Cap fund onethat is supposed to invest in large-cap stocks10

We augment the regression specification ofTable 3 by introducing a dummy variableIndLC (which equals one if a fund is a Large

Cap fund and zero otherwise) and an addi-tional interaction term involving LOGTNA andIndLC as in equation (5) We first report theresults for gross fund returns The coefficient infront of LOGTNA is about 003 (across thefour performance benchmarks) Importantlythe coefficient in front of the interaction term ispositive and statistically significant (about 003across the four performance benchmarks) Thisis the sign predicted by the liquidity hypothesissince it says that for Large Cap funds there isno effect of fund size on performance Theresults using net fund returns reported aresimilar

These findings suggest that liquidity plays animportant role in eroding performance More-over as many practitioners have pointed outsince managers of funds get compensated onassets under management they are not likelyvoluntarily to keep their funds small just be-cause it hurts the returns of their investors whomay not be aware of the downside of scale (seeBecker and Vaughn 2001 and Section IV forfurther discussion)11

As we pointed out in the beginning of thepaper however liquidity means that largefunds need to find more stock ideas than smallfunds but it does not therefore follow thatthey cannot Indeed large funds can hiremore managers to follow more stocks To seethat this is possible we calculate some basicsummary statistics on fund holdings by fundsize quintiles Since the CRSP Mutual FundDatabase does not have this information weturn to the CDA Spectrum Database We takedata from the end of September 1997 andcalculate the number of stocks held by eachfund The median fund in the smallest fundsize quintile has about 16 stocks in its port-folio while the median fund in the largestfund size quintile has about 66 stocks in itsportfolio even though the large funds aremany times bigger than their smaller counter-parts These numbers make clear that largefunds do not significantly scale up the numberof stocks that they hold or cover relative totheir smaller counterparts Yet there is plenty

10 We merged the CRSP Mutual Fund Database and theCDA Spectrum Database so that we have information on thestocks held by the funds in our sample We have verifiedthat mutual funds with the self-reported fund style of SmallCap Growth do indeed invest in stocks with a much lowermarket capitalization than held by funds of other styles Incontrast mutual funds whose self-reported fund style isldquoGrowth and Incomerdquo invest in much larger stocks whencompared to funds with other styles As a result we char-acterize these funds as Large Cap funds in our analysisabove Moreover we have also verified that Small Capfunds hold stocks that are more illiquid in the sense that thestocks in their portfolios tend to have much larger Kylelambdas and bid-ask spreads than the stocks held byfunds of other styles And Large Cap funds hold stocksthat are much more liquid than those held by funds ofother styles

11 Related literature finds that mutual fund investors aresusceptible to marketing (see eg Gruber 1996 Eric RSirri and Peter Tufano 1998 Lu Zheng 1999)


TABLE 5mdashEFFECT OF FUND SIZE ON PERFORMANCE BY FUND STYLE

Panel A Small Cap Growth funds



INTERCEPT 0205 0257 0320 0263 0175 0226 0289 0231(127) (153) (244) (198) (108) (134) (220) (175)

INDnot SCG 0250 0252 0260 0262 0249 0251 0259 0261(154) (155) (162) (163) (153) (154) (161) (162)

LOGTNAit1 0058 0058 0063 0061 0055 0056 0060 0058(315) (317) (340) (329) (303) (305) (328) (317)

LOGTNAit1 0040 0041 0043 0044 0040 0041 0043 0044INDnot SCG (236) (239) (253) (256) (235) (237) (252) (255)

LOGFAMSIZEit1 0012 0012 0011 0011 0012 0012 0012 0012(257) (257) (254) (254) (259) (259) (257) (256)

TURNOVERit1 0000 0000 0000 0000 0000 0000 0000 0000(162) (161) (162) (161) (157) (156) (157) (157)

AGEit1 0001 0001 0001 0001 0001 0001 0001 0001(098) (097) (098) (097) (087) (087) (088) (087)

EXPRATIOit1 0018 0018 0019 0019 0062 0062 0064 0063(044) (043) (046) (044) (149) (148) (152) (151)

TOTLOADit1 0001 0001 0001 0001 0001 0001 0001 0001(054) (054) (052) (052) (048) (047) (046) (046)

FLOWit1 0000 0000 0000 0000 0000 0000 0000 0000(096) (095) (090) (092) (102) (102) (096) (099)

FUNDRETit1 0022 0022 0021 0021 0022 0022 0022 0022(345) (345) (343) (342) (348) (348) (345) (345)

No of months 228 228 228 228 228 228 228 228

Panel B Large Cap funds



INTERCEPT 0016 0068 0128 0068 0017 0035 0095 0034(019) (084) (172) (092) (020) (043) (127) (047)

INDLC 0119 0121 0119 0120 0112 0114 0112 0113(131) (133) (133) (134) (123) (125) (124) (126)

LOGTNAit1 0029 0029 0032 0029 0027 0027 0029 0027(222) (224) (245) (225) (202) (205) (226) (205)

LOGTNAit1 0031 0031 0031 0031 0030 0030 0030 0030INDLC (229) (232) (234) (236) (223) (226) (228) (230)

LOGFAMSIZEit1 0011 0011 0011 0011 0011 0011 0011 0011(254) (254) (251) (251) (258) (258) (255) (255)

TURNOVERit1 0000 0000 0000 0000 0000 0000 0000 0000(180) (179) (180) (179) (175) (175) (176) (175)

AGEit1 0001 0001 0001 0001 0001 0001 0001 0001(126) (125) (126) (125) (117) (117) (117) (116)

EXPRATIOit1 0020 0019 0021 0020 0065 0065 0066 0065(049) (048) (052) (050) (163) (162) (166) (164)

TOTLOADit1 0001 0001 0001 0001 0001 0001 0001 0001(035) (035) (034) (035) (030) (030) (029) (029)

FLOWit1 0000 0000 0000 0000 0000 0000 0000 0000(084) (082) (077) (079) (089) (088) (083) (085)

FUNDRETit1 0020 0020 0020 0020 0021 0020 0020 0020(310) (309) (307) (307) (313) (312) (310) (309)

No of months 228 228 228 228 228 228 228 228

Notes This table reports the Fama-MacBeth (1973) estimates of monthly fund returns regressed on fund characteristics lagged one month Thesample includes only funds that fall within fund size quintiles two to five Fund returns are calculated before (gross) and after (net) deductingfees and expenses These fund returns are adjusted using the market model the CAPM the Fama-French (1993) 3-Factor model and the4-Factor model In panel (A) the regression specification is the one in Table 3 but augmented with INDnot SCG which is a dummy variablethat equals one if the self-reported fund style is not Small Cap Growth and zero otherwise and this indicator variable interacted with LOGTNAIn panel (B) instead of INDnot SCG we augment the regression with INDLC which is a dummy variable that equals one if the fund styleis identified as a Large Cap fund The sample is from January 1981 to December 1999 The t-statistics are adjusted for serial correlation usingNewey-West (1987) lags of order three and are shown in parentheses


of scope for them to do so given the thousandsof stocks available

C The Role of Organization The Effect ofFamily Size on Performance

To see that assets under management neednot be bad for the performance of a fund orga-nization recall from Table 3 that controlling forfund size assets under management of the otherfunds in the family that the fund belongs toactually increase the fundrsquos performance Thecoefficient in front of LOGFAMSIZE is roughly0007 regardless of the performance benchmarkused One standard deviation of this variable is275 so a two-standard deviation shock in thesize of the family that the fund belongs to leads

to about a 385-basis-point movement in thefundrsquos performance the following month (orabout a 46-basis-point movement annually) de-pending on the performance measure used Theeffect is smaller than that of fund size on returnsbut is nonetheless statistically and economicallysignificant In other words assets under man-agement are not bad for a fund organizationrsquosperformance per se

In Table 6 we extend our analysis of theeffect of family size on fund returns by seeingwhether this effect varies across fund stylesOur hope is that family size is just as importantfor Small Cap Growth funds as for other fundsAfter all it is these funds that are most affectedby scale For us to claim that scale is not bad perse even accounting for liquidity we would like

TABLE 6mdashEFFECT OF FAMILY SIZE ON PERFORMANCE BY FUND STYLE

Panel A Fund size quintiles two to five



LOGTNAit1 0051 0052 0056 0054 0048 0049 0053 0051(283) (286) (313) (302) (270) (272) (301) (289)


LOGFAMSIZEit1 0007 0007 0007 0007 0007 0007 0007 0007(070) (070) (069) (070) (070) (070) (069) (070)

LOGFAMSIZEit1 0004 0004 0004 0004 0004 0004 0004 0004INDnot SCG (041) (041) (041) (040) (042) (042) (042) (041)

No of months 228 228 228 228 228 228 228 228

Panel B All fund size quintiles one to five



LOGTNAit1 0067 0070 0071 0072 0064 0067 0068 0069(367) (383) (387) (391) (351) (367) (371) (376)


LOGFAMSIZEit1 0012 0012 0012 0012 0012 0012 0012 0012(113) (113) (111) (111) (115) (115) (113) (113)


No of months 228 228 228 228 228 228 228 228

Notes This table reports the Fama-MacBeth (1973) estimates of monthly fund returns regressed on fund characteristics laggedone month Fund returns are calculated before (gross) and after (net) deducting fees and expenses These returns are adjustedusing the market model the CAPM the Fama-French (1993) 3-Factor model and the 4-Factor model The regressionspecification is the one in Table 5 but augmented with Indnot SCG which is a dummy variable which equals one if theself-reported fund style is Not Small Cap Growth and zero otherwise interacted with LOGFAMSIZE Estimates of theintercept and the other independent variables are omitted for brevity The other independent variables include TURNOVERAGE EXPRATIO TOTLOAD FLOW and LAGFUNDRET The sample is from January 1981 to December 1999 Thet-statistics are adjusted for serial correlations using Newey-West (1987) and are shown in parentheses


to find that the benefits of family size are de-rived even by funds that are most affected byliquidity To see whether this is the case weaugment the regression specification in Table5 by adding an interaction term involving LOG-FAMSIZE and the dummy indicator Indnot SCG(which equals one if a fund is not Small CapGrowth and zero otherwise) The variable ofinterest is the coefficient in front of this inter-action term

Panel (A) presents the regression results forfunds in fund size quintiles two through fiveThe coefficient in front of this interaction termis positive but is not statistically significant Inother words it does not appear that there aremajor differences between the effect of familysize on performance among Small Cap Growthfunds and other funds This finding is consistentwith our story To make sure that this findingholds more generally we reestimate this regres-sion specification using all funds The coeffi-cient in front of the interaction term is negativebut is again not statistically significant Usingeither sample it does not appear that the effectof family size is limited to Not Small CapGrowth funds So we can be assured that fundsdo benefit from being part of a large family

This findingmdashfund performance declineswith own fund size but increases with the size ofthe other funds in the familymdashfits nicely withanecdotal evidence from industry practitioners(see eg Pozen 1998 Albert J Fredman andRuss Wiles 1998) One plausible interpretationof the family size finding is that it is capturingeconomies of scale associated with marketingThis begs the question however of why afundrsquos expense ratio is not capturing this scaleeffect since a fundrsquos expense ratio is on theright-hand side of the cross-sectional fund re-turn regressions

It turns out that the expense ratio reported bymutual funds accounts for only managementadministrative and marketing fees12 Tradingcommissions charged by brokers and lendingfees for (shorting) stocks are not treated as partof expenses but are simply deducted or added toincome and hence net returns It is well knownthat there are tremendous economies of scale

associated with trading commissions and lend-ing fees at the family level Bigger families likeFidelity are able to get better concessions ontrading commissions and earn higher lendingfees for the stocks held by their funds

It is difficult to figure out these trading com-missions because families bury them in theirfund prospectuses Nonetheless practitionersbelieve these costs are substantial and can be asmuch as a few percentage points a year Hencethe spread in performance between funds be-longing to large versus those in small familiescan easily be accounted for by the economies ofscale in trading commissions and lending fees atthe family level13

Importantly according to industry anecdotesin most families major decisions are decentral-ized in that the fund managers make stock pickswithout substantial coordination with othermanagers Managers in charge of differentfunds choose stocks as they see fit without wor-rying about resources being taken away fromthem by the family So a family is an organiza-tion that credibly commits to letting each of itsfund managers run his own assets

As such the family size finding makes clearthat liquidity and scale need not be bad for fundperformance depending on how the fund isorganized After all if a large fund is organizedlike a fund family with a number of small fundsrun by different managers then scale need notbe bad per se just as family size does not appearto be bad for family performance

More importantly the finding regarding fam-ily size also helps us further to rule out alterna-tive hypotheses related to managers of largefunds not caring about maximizing net returnsThis line of argument does not likely supportour findings since managers in large familiesapparently do care enough about net fund re-turns to make sufficient investments to maintainperformance at the family level Moreovermost of the extant evidence on fund flows indi-cates that managers care about net fund returns

12 See John Hechinger (2004) for a detailed discussion ofmutual fund expense accounting

13 Zoran IvKovic (2002) argues that being part of a largefamily may improve performance because of other spill-overs In his analysis of family size he happened to controlfor fund size and found that fund size indeed erodes per-formance Though the goal of his paper was not to examinefund size it is comforting to know that some of our baselineresults have been independently verified


since better past returns lead to larger assetsunder management (see eg Sirri and Tufano1998 Anthony W Lynch and David Musto2003)

D Organizational Diseconomies and FundPerformance

If managers care about net fund returns andscale need not be bad for performance per sewhy then does it appear that scale erodes fundperformance because of liquidity We hypoth-esize that in addition to liquidity fund sizeerodes performance because of organizationaldiseconomies To make things concrete imag-ine that there is a small fund company X withone fund operated by one manager who picksthe stocks Since the fund is small the managercan easily invest the assets under managementby generating a few stock ideas Now imaginethat there is a large fund Y in which the managerno longer has the capacity to invest all themoney So the manager needs co-managers tohelp him run the fund For fund Y the stockpicks need to be coordinated among many moreagents and therefore organizational form (egflat versus hierarchical) becomes important Soorganizational diseconomies may arise

There are many types of organizational dis-economies that lead to different predictions onwhy small organizations outperform large onesOne set of diseconomies from the work ofOliver E Williamson (1975 1988) includesbureaucracy and related coordination costsAnother set of diseconomies comes from theinfluence-cost literature (see eg Paul R Mil-grom and Roberts 1988) Yet another set ofdiseconomies centers on the adverse effects ofhierarchies (or authority) for the incentives ofagents who do not have any control over asset-allocation decisions (see eg Aghion and Ti-role 1997 Stein 2002)

Interestingly the findings in Tables 3 and 6allow us to discriminate among different typesof organizational diseconomies For instance ifWilliamsonian diseconomies are behind the re-lationship between size and performance thenone expects that funds that belong to large fam-ilies do worse since bureaucracy ought to bemore important in huge fund complexes Thefact that we do not find this indicates that bu-reaucracy is not likely an important reason

behind why performance declines with fundsize

We conjecture that hierarchy costs may beespecially relevant for mutual funds We take acloser look at the effect of organizational dis-economies due to hierarchy costs on fund per-formance by testing some predictions fromStein (2002) Stein argues that in the presenceof such hierarchy costs small organizationsought to outperform large ones at tasks thatinvolve the processing of soft information (ieinformation that cannot be directly verified byanyone other than the agent who produces it)The basic premise is that in larger organizationswith hierarchies the process of agents fightingto have their ideas implemented will affect out-comes If the information is soft then agentshave a hard time convincing others of theirideas and it is more difficult to pass this infor-mation up the organization14

In the context of mutual funds this meansthat in funds in which managers fight to havetheir ideas implemented efforts to uncover cer-tain investment ideas in this setting are dimin-ished relative to a situation in which themanagers control their own smaller funds Forinstance managers may end up expending toomuch research effort on quantitative measuresof a company (ie hard information) so as toconvince others to implement their ideas eventhough more time and effort might ideally beinvested in talking to CEOs of companies andgetting an impression of them (ie soft infor-mation) All else equal large funds may per-form worse than small ones

Note that one of our previous findingsmdashthatfund performance does not decline with familysizemdashis consistent with Stein (2002) since hi-

14 Steinrsquos analysis also indicates that large organizationsmay actually be very efficient at processing hard informa-tion In the context of the mutual fund industry one canthink of passive funds that mimic indices as primarilyprocessing hard information As such one would expectthese types of funds not to be very much affected by scaleTwo pieces of evidence are consistent with this predictionFirst Vanguard seems to dominate the business for indexfunds Indeed for indexers being large is an asset since onecan then acquire better tracking technologies and bettercomputer programmers Second our evidence in Table5 that fund size affects only small cap funds is consistentwith this observation since most index funds which tend tomimic the SampP 500 index trade predominantly large stocks


erarchy costs are not present at the family levelprecisely because the family typically agreesnot to reallocate resources across funds Differ-ent funds in a family have their own boards thatdeal with such issues as managerial replace-ment So the manager in charge of a fund gen-erally does not have to worry about the familytaking away his resources and giving those re-sources to some other fund in the family

Below we test several predictions of Stein(2002) to see if the inverse relationship betweenfund size and performance is indeed influencedby organizational diseconomies

Fund Size and Fund InvestmentsmdashOne pre-diction of Stein (2002) is that we expect thatsmall funds are better than large ones which aremore likely to have hierarchies at the process-ing of soft information To test this predictionwe compare the investments of large and smallfunds in local stocks (companies located nearwhere a fund is headquartered) Investing insuch companies requires that the organizationprocess soft information as opposed to a strictlyquantitative investing approach which wouldtypically process hard information like price-to-earnings ratios

Our work in testing this prediction builds onthe very interesting work of Coval and Mos-kowitz (1999 2001) whose central thesis is thatmutual fund managers do have ability when itcomes to local stocks Anecdotal evidence indi-cates that this ability comes in the form ofprocessing soft information eg talking to andevaluating CEOs of local companies They findthat funds can earn superior returns on theirlocal investments We focus on the effect of sizeon investment among small cap funds We arealso interested in the effect of family size on thecomposition of fund investments

The CDA Spectrum Database does not havethe same information on fund characteristics asthe CRSP Mutual Fund Database Luckily weare able to construct from the fund stock hold-ings data a proxy for fund size which is simplythe value of the fundrsquos stock portfolio at the endof a quarter While this is not exactly the sameas asset base since funds hold cash and bondswe believe that the proxy is a reasonable onebecause we can look at the tails of the sizedistribution to draw inferences ie comparevery small funds to other funds Moreover the

noise in our fund size measure does not obvi-ously bias our estimates

In addition we can construct better style con-trols for each fund by looking at their stockholdings We use a style measure constructedby Daniel et al (1997) which we call theDGTW style adjustment For each month eachstock in our sample is characterized by its lo-cation in the (across-stocks) size quintilesbook-to-market quintiles and price-momentumquintiles So a stock in the bottom of the sizebook-to-market and momentum quintiles in aparticular month would be characterized by atriplet (1 1 1) For each fund we can charac-terize its style along the size book-to-marketand momentum dimension by taking the aver-age of the DGTW characteristics of the stocksin their portfolio weighted by the percentage ofthe value of their portfolio that they devote toeach stock We can then define a small cap fundas one whose DGTW size measure falls in thebottom 10 percent when compared to otherfunds

Table 7 reports the effect of fund size on thepercentage of the value of a fundrsquos portfolio

TABLE 7mdashEFFECT OF FUND AND FAMILY SIZE ON

INVESTMENTS IN LOCAL STOCKS

(1) (2) (3)

SMALLFUND 00759 00580 00584(949) (674) (679)

SMALLCAPFUND 00515 00340 00341(628) (386) (388)

SMALLFUND 01244 01229SMALLCAPFUND (558) (580)

FAMILYSIZE 00001(101)

Momentum effects Yes Yes YesBook-to-market effects Yes Yes YesFund city effects Yes Yes Yes

Notes This table shows the effect of fund and family sizeon mutual fund investments in local stocks The dependentvariable is the percentage of the value of a fundrsquos portfolioinvested in local stocks SMALLFUND equals one if afundrsquos size is in the bottom 10 percent of the fund sizedistribution and zero otherwise SMALLCAPFUND equalsone if a fundrsquos Daniel Grinblatt Titman and Wermers(1997) small cap style score is in the bottom 10 percentacross funds FAMILYSIZE is the number of funds in thefamily that the fund belongs to Momentum effects andbook-to-market effects control for other differences in fundstyles Data are from the end of September 1997 Robustt-statistics are reported in parentheses


devoted to local stocks where locality is de-fined relative to the headquarters of the fundsas in Coval and Moskowitz (1999 2001) Spe-cifically a stock is considered a local fundinvestment if the headquarters of the companyis in the same census region as the mutualfundrsquos headquarters15 The dependent variableis the total value of the local stocks in the fundportfolio divided by the total value of the stocksin the fund portfolio The first independent vari-able is SMALLFUND which is a dummy vari-able that equals one if the fund is in the bottom10 percent of the fund size distribution SMALL-CAPFUND equals one if a fund has an averageDGTW size score for its stocks that is in thebottom 10 percent across funds The regressionsalways have Fund City Effects (ie the citywhere the fund is headquartered) as controls

From column (1) small funds are more likelyto invest a larger percentage of their portfolio inlocal stocks than do small cap funds The aver-age percentage of stocks that are local in afundsrsquo portfolio is 1657 percent The coeffi-cient on SMALLFUND is 00759 That meansthat being a small fund increases the amount oflocal stocks a fund owns by 00759 divided by01657 or 46 percent In column (2) the coef-ficient of interest is the interaction term involv-ing small and small cap funds It is positive andstatistically significant which is consistent withour conjecture The coefficient of 01244 meansthat being a small fund and a small cap fundraises the percentage of local stock investmentson average by 01244 divided by 01657 or 75percent

One interesting question is the degree towhich family size affects a fundrsquos investmentpolicy In column (3) we add a family-sizeproxy as measured by the number of funds inthe family The coefficient in front of this vari-able is positive but the economic magnitude issmall and it attracts a t-statistic of only one Inother words family size does not have an effect

on whether a fund invests in local stocks Ifwe define family size as the log of the assets ofthe other funds in the fundrsquos family then weget similarly small economic magnitudes andt-statistics near zero For brevity we omit theseadditional results Again one can interpret thisresult as consistent with Steinrsquos model to theextent that fund size has an effect on fund actionbut not family size

These findings however may also be consis-tent with larger funds that have more resourcesto visit companies not located nearby To dis-tinguish between this explanation and our pre-ferred explanation that small funds and fundsbelonging to larger families are better at theprocessing of soft information we see in Table8 whether the local stocks picked by small fundsdo better than those picked by large funds Thedependent variable is the return to the fundrsquosinvestments in local stocks In column (1) wefind that small funds and small cap funds notonly invest more of their assets in local stocksbut also do better at investing in them Thestandard deviation of local stock investmentreturns in our sample is 02077 Being a small

15 There are nine census regions New England (CTMA ME NH RI VT) Middle Atlantic (NJ NY PA) EastNorth Central (IL IN MI OH WI) West North Central(IA KS MN MO NE ND SD) South Atlantic (DE FLGA MD NC SC VA WV) East South Central (AL KYMS TN) West South Central (AR LA OK TX) Moun-tain (AZ CO ID MT NE NM UT WY) and Pacific(AK CA HI OR WA)

TABLE 8mdashEFFECT OF FUND AND FAMILY SIZE ON THE

PERFORMANCE OF INVESTMENTS IN LOCAL STOCKS

(1) (2) (3)

SMALLFUND 00258 00143 00150(183) (094) (099)

SMALLCAPFUND 01190 01078 01079(821) (695) (696)



Momentum effects Yes Yes YesBook to market effects Yes Yes YesFund city effects Yes Yes Yes

Notes This table shows the effect of fund and family sizeon the performance of mutual fund investments in localstocks The dependent variable is the 4-quarter return tofund investments in local stocks SMALLFUND equals oneif a fundrsquos size is in the bottom 10 percent of the fund sizedistribution and zero otherwise SMALLCAPFUND equalsone if a fundrsquos Daniel Grinblatt Titman and Wermers(1997) small cap style score is in the bottom 10 percentacross funds FAMILYSIZE is the number of funds in thefamily that the fund belongs to Momentum effects andbook-to-market effects control for other differences in fundstyles Data are from the end of September 1997 Robustt-statistics are reported in parentheses


fund raises local return by 00258 divided by02077 or 12 percent of a standard deviation Incolumn (2) we add an interaction term involv-ing these two variables and find that the perfor-mance difference between small funds and otherfunds is especially big in small cap funds Incolumn (3) we add a family size proxy asmeasured by the number of funds in the familyThe coefficient is positive but is statisticallyinsignificant So a fund that belongs to a largefamily does not do better in local stock invest-ments Similar results obtain using total netassets for the family

The findings in Table 8 suggest that the per-formance differences between small and largefunds have something to do with their differingability to invest in local stocks Note that Tables7 and 8 report the results using only the cross-section in September 1997 We have replicatedour analysis using the other quarters in the pe-riod of 1997 to 1998 The results are very robustacross quarters

Management Structure and the Compositionof Fund InvestmentsmdashPerhaps a more directimplication of Steinrsquos model is to look at therelationship between the management structureof a fund and the fundrsquos investments When afund is co-managed and there is more fighting toimplement ideas then we should find thatcontrolling for fund size funds run by commit-tee should invest less in local stocks (ie man-agers should pitch hard-information stockideas) and do worse in these soft-informationstocks

To test this prediction we obtain data on thenumber of managers running a fund for thecross-section of funds described in Tables 7 and8 In the CRSP database there is a field that liststhe names of managers in charge of pickingstocks We create a new variable MT whichcaptures how many managers are running thefund If there is only one name then MT equalsone If there are two names then MT equalstwo and so forth The maximum number ofnames in the database is four Moreover a smallfraction of the funds are listed as ldquoTeam Man-agedrdquo For these funds we set MT equal to fouras well Thus MT can take on values of onetwo three or four From reading the prospec-tuses of various funds it appears that funds thathave co-managers do make decisions within a

committee We lose a fraction of the cross-section of funds in Tables 7 and 8 because somefunds do not have entries in the manager field

In Table 9 we consider the effect of mana-gerial structure on the composition of fund in-vestments in local stocks and the performanceof these local investments The first column ofTable 9 is similar to those in Table 7 exceptthat we have added a dummy variable MULTI-MANAGER which equals one when MT isgreater than one and is zero otherwise asan additional explanatory variable Moreoversince we are concerned with isolating the effectof managerial structure we have introducedinto this regression very conservative fund sizefamily size and style controls Even with theseconservative controls the coefficient in front ofMULTI-MANAGER is 0012 with a t-statisticof 175 In other words a solo-managed fund issignificantly more likely to invest in localstocks than team-managed funds The economicmagnitude is smaller than that of being a smallfund but is still economically interesting In thissample the average percentage of stocks thatare local in a fundrsquos portfolio is 152 percentSince the coefficient on MULTI-MANAGER is

TABLE 9mdashEFFECT OF MANAGEMENT STRUCTURE ON THE

AMOUNT AND PERFORMANCE OF INVESTMENTS IN LOCAL

STOCKS

Holdings Returns(1) (2)

MULTI-MANAGER 00102 00190(175) (173)

Fundsize effects Yes YesFamily size effects Yes YesMomentum effects Yes YesBook-to-market effects Yes YesStyle effects Yes YesFund city effects Yes Yes

Notes This table shows the effect of fund managementstructure on the holdings and performance of mutual fundinvestments in local stocks The dependent variable in col-umn (1) is the percentage of the value of a fundrsquos portfolioinvested in local stocks The dependent variable in column(2) is the 4-quarter return-to-fund investments in localstocks MULTI-MANAGER is an indicator that the fund ismanaged by more than one person Fundsize effects familysize effects momentum effects book-to-market effects andstyle effects control for other differences in fund stylesFund city effects are indicators for the city where the fundis located Data are from the end of September 1997 Robustt-statistics are reported in parentheses


0012 being a committee-managed fund de-creases the amount of local stocks a fund ownsby 0012 divided by 0152 or about 8 percent

The second column of Table 9 is similar tothose in Table 8 except that we have added thevariable MULTI-MANAGER as an additionalexplanatory variable for the performance offund local investments Again we have intro-duced into this regression very conservativefund size family size and style controls Thecoefficient in front of MULTI-MANAGER is00190 with a t-statistic of 173 suggestingthat co-managed funds also do worse at pickinglocal companies The standard deviation of lo-cal stock investment returns in this sample is0186 So being a committee-managed fund de-creases local return by 00190 divided by 0186or 10 percent of a standard deviation Bothfindings are consistent with Steinrsquos model

Management Structure and Fund Perfor-mancemdashFinally we look to see whether solo-managed funds outperform team-managed onesafter controlling for fund size Our prior is thatorganization form might be a less powerful pre-dictor of performance than size The reason isthat few managers would turn away moremoney even if it hurt net fund returns for inves-tors But conditional on a certain fund size onemight think that the fund would choose thebetter organizational form to maximize net fundreturns In other words conditional on sizeform may be a less powerful predictor of per-formance Moreover we are much more limitedin terms of data when looking at the effect ofmanagerial structure on fund performanceWhereas we are able to obtain data on fund sizeback to the 1960s we are able to obtain data onmanagerial structure only back to 1992 becausethe data on how many managers are running afund are unavailable from the CRSP before thatdate As before we create the MT variablefor as many funds as possible back to 1992Despite these caveats it is interesting to lookat the effect of managerial structure on fundperformance

The result for the effect of management struc-ture on fund returns is given in Table 10 Theupshot is that all the previous results regardingfund size continue to hold For instance thecoefficient in front of fund size is the same asin Table 3 We find that funds run by com-

mittee underperform by about 4 basis points amonth or 48 basis points a year This effect iseconomically smaller than the effect of fundsize but is statistically significant Again thisfinding on performance is consistent withSteinrsquos model

IV Alternative Explanations and FurtherDiscussion

In Section III we ruled out a number ofalternative explanations for our findings in es-tablishing the role of liquidity and organizationin influencing the relationship between fundsize and performance In this section we dis-cuss additional explanations One institutionalreason for our finding may be that it is easier forfund families to manipulate the performance ofsmall funds in small cap stocks While thisalternative explanation is possible we do notthink that it is driving our results for severalreasons First we exclude the very smallestfunds for which this is likely and our results areunchanged Moreover in analysis that is notreported for reasons of brevity we have alsodropped out very young funds for which suchmanipulation is most likely and our results areagain unchanged Finally such manipulationcannot account for the relationship betweenfund size and the composition of fund invest-ments or for the relationship between manage-rial structure and fund investments andperformance documented in Section III D

Besides such institutional reasons it may bethe case that our findings are due to other orga-nization-related issues as well as the hierarchy-cost hypothesis presented in Section III D Forinstance many managers talk about how diffi-cult it is to train new hires when their fundgrows We view such comments as broadlysupportive of our contention that organizationmatters for fund performance

Ideally we would like to have informationabout the incentives inside fund organizationsFor instance our hierarchy-cost hypothesis alsosuggests that a crucial unobservable determi-nant of fund performance is the nature of theincentives inside the fund More concretelysuppose that the organizational diseconomiesare due to the adverse effects of a hierarchy onmanagerial effort Then an optimal organiza-tional structure is to limit managers to a small


pot of money and let them manage it as theychoose With such an organizational structurescale may not affect performance This is ex-actly what happens at the family level and prob-ably why performance does not decline withfamily size

Our analysis here is similar to the work ofAllen N Berger et al (2002) They test the ideain Stein (2002) that small organizations are bet-ter than large ones in activities that require theprocessing of soft information in the context ofbank lending to small firms They find that largebanks are less willing than small ones to lend tofirms that do not keep formal financial recordsThey also find that large banks lend at a greaterdistance interact more impersonally with theirborrowers have shorter and less exclusive rela-tionships and do not alleviate credit constraintsas effectively Interestingly they find that thesize of the bank-holding company that a bankbelongs to does not affect the lending policies ofthe bank In other words size itself is not nec-

essarily bad for banks This finding is related toour finding regarding the neutral-to-beneficialeffects of family size on fund performanceThey argue that their findings are most con-sistent with the property-rights approach ofGrossman-Hart-Moore regarding the disadvan-tages of integration

Our paper is also related to other papers thatattempt to test the basic Grossman-Hart-Mooreinsight in particular settings Notable examplesinclude George P Baker and Thomas Hubbard(2000) whose work centers on the truckingindustry and the question of whether driversshould own the trucks they operate and DuncanSimester and Binger Wernerfelt (2000) wholook at the ownership of tools in the carpentryindustry

V Conclusion

To the best of our knowledge we are the firstto find strong evidence that fund size erodes

TABLE 10mdashEFFECT OF FUND MANAGEMENT STRUCTURE ON PERFORMANCE



INTERCEPT 0069 0018 0039 0008 0092 0042 0014 0032(066) (019) (025) (005) (088) (044) (009) (021)

MULTI-MANAGERi 0040 0040 0041 0041 0040 0041 0042 0042(222) (223) (230) (230) (226) (227) (234) (234)

LOGTNAit1 0024 0023 0026 0025 0022 0022 0025 0023(234) (230) (198) (183) (219) (215) (186) (170)

LOGFAMSIZEit1 0016 0016 0016 0016 0016 0016 0016 0016(238) (238) (238) (238) (240) (239) (240) (239)

TURNOVERit1 0000 0000 0000 0000 0000 0000 0000 0000(128) (129) (128) (129) (126) (126) (126) (126)

AGEit1 0001 0001 0001 0001 0001 0001 0001 0001(064) (065) (064) (064) (060) (061) (059) (059)

EXPRATIOit1 0002 0002 0003 0003 0043 0043 0045 0045(003) (003) (004) (004) (054) (054) (055) (055)

TOTLOADit1 0002 0002 0002 0002 0002 0002 0002 0002(028) (029) (030) (029) (036) (036) (037) (037)

FLOWit1 0000 0000 0000 0000 0000 0000 0000 0000(164) (168) (168) (169) (164) (168) (168) (169)

FUNDRETit1 0030 0030 0029 0029 0031 0030 0030 0030(338) (338) (334) (334) (341) (341) (337) (337)

No of months 96 96 96 96 96 96 96 96

Notes This table reports the Fama-MacBeth (1973) estimates of monthly fund returns regressed on fund characteristics laggedone month The sample includes only funds that fall within fund size quintiles two to five Fund returns are calculated before(gross) and after (net) deducting fees and expenses These returns are adjusted using the market model the CAPM theFama-French (1993) 3-Factor model and the 4-Factor model The regression specification is the one in Table 3 but augmentedwith MULTI-MANAGERi which is a dummy variable that equals one if the fund is managed by two or more individuals orby a team The sample is from January 1992 to December 1999 The t-statistics are adjusted for serial correlation usingNewey-West (1987) lags of order three and are shown in parentheses


performance We then consider various expla-nations for why this might be the case We findthat this relationship is not driven by heteroge-neity in fund styles fund size being correlatedwith other observable fund characteristics orany type of survivorship bias Instead we findthat the effect of fund size on fund returns ismost pronounced for funds that play small capstocks This suggests that liquidity is an impor-tant reason why size erodes performance More-over we argue that organizational diseconomiesrelated to hierarchy costs may also play a role inaddition to liquidity in the documented disec-onomies of scale Consistent with this view wefind that the size of a fundrsquos family does notsignificantly erode fund performance Finallyusing data on whether funds are solo-managedor co-managed and the composition of fundinvestments we find that organizational disec-onomies affect the relationship between fundsize and performance along the lines predictedby Stein (2002)

Importantly our findings have relevancefor the ongoing research into the question ofRonald H Coase (1937) regarding the determi-nants of the boundaries of the firm While anenormous amount of theoretical research hasbeen done on this question there has been farless empirical work Our findings suggest thatmutual funds may be an invaluable laboratorywith which to study related issues in organiza-tion The amount of data on mutual funds un-like that on corporations is plentiful because ofdisclosure regulations and the fact that the tasksand performance of mutual funds are measur-able We plan to refine our understanding of thenature of organizations in the future using thislaboratory

REFERENCES

Aghion Philippe and Tirole Jean ldquoFormal andReal Authority in Organizationsrdquo Journalof Political Economy 1997 105(1) pp 1ndash29

Almazan Andres Brown Keith C MurrayCarlson and Chapman David A ldquoWhy Con-strain Your Mutual Fund Managerrdquo Unpub-lished Paper 2001

Baker George P and Hubbard Thomas N ldquoCon-tractibility and Asset Ownership OnboardComputers and Governance in US Truckingrdquo

National Bureau of Economic Research IncNBER Working Papers 7634 2000

Becker Stan and Vaughan Greg ldquoSmall IsBeautifulrdquo Journal of Portfolio Manage-ment 2001 27(4) pp 9ndash18

Berger Allen N Miller Nathan H PetersenMitchell A Rajan Raghuram G and SteinJeremy C ldquoDoes Function Follow Organiza-tional Form Evidence from the LendingPractices of Large and Small Banksrdquo Na-tional Bureau of Economic Research IncNBER Working Papers 8752 2002

Berk Jonathan B and Green Richard C ldquoMu-tual Fund Flows and Performance in RationalMarketsrdquo National Bureau of Economic Re-search Inc NBER Working Papers 92752002

Bolton Patrick and Scharfstein David S ldquoCor-porate Finance the Theory of the Firm andOrganizationsrdquo Journal of Economic Per-spectives 1998 12(4) pp 95ndash114

Brown Keith C Harlow W V and StarksLaura T ldquoOf Tournaments and TemptationsAn Analysis of Managerial Incentives in theMutual Fund Industryrdquo Journal of Finance1996 51(1) pp 85ndash110

Carhart Mark M ldquoOn Persistence in MutualFund Performancerdquo Journal of Finance1997 52(1) pp 57ndash82

Chen Joseph Hong Harrison and Stein JeremyC ldquoBreadth of Ownership and Stock Re-turnsrdquo Journal of Financial Economics2002 66(2ndash3) pp 171ndash205

Chevalier Judith and Ellison Glenn ldquoRisk Tak-ing by Mutual Funds as a Response to Incen-tivesrdquo Journal of Political Economy 1997105(6) pp 1167ndash200

Chevalier Judith and Ellison Glenn ldquoCareerConcerns of Mutual Fund Managersrdquo Quar-terly Journal of Economics 1999 114(2) pp389ndash432

Coase Ronald H ldquoThe Nature of the FirmrdquoEconomica 1937 4 pp 386ndash405

Coval Joshua D and Moskowitz Tobias JldquoHome Bias at Home Local Equity Prefer-ence in Domestic Portfoliosrdquo Journal of Fi-nance 1999 54(6) pp 2045ndash73

Coval Joshua D and Moskowitz Tobias J ldquoTheGeography of Investment Informed Tradingand Asset Pricesrdquo Journal of Political Econ-omy 2001 109(4) pp 811ndash41

Daniel Kent Grinblatt Mark Titman Sheri-


dan and Wermers Russ ldquoMeasuring MutualFund Performance with Characteristic-Based Benchmarksrdquo Journal of Finance1997 52(3) pp 1035ndash58

Elton Edwin J Gruber Martin J and BlakeChristopher R ldquoA First Look at the Accuracyof the CRSP Mutual Database and a Com-parison of the CRSP and Morningstar MutualFund Databasesrdquo Journal of Finance 200156(6) pp 2415ndash30

Falkenstein Eric G ldquoPreferences for StockCharacteristics as Revealed by Mutual FundPortfolio Holdingsrdquo Journal of Finance1996 51(1) pp 111ndash35

Fama Eugene F and French Kenneth R ldquoCom-mon Risk Factors in the Returns on Stocksand Bondsrdquo Journal of Financial Econom-ics 1993 33(1) pp 3ndash56

Fama Eugene F and MacBeth James D ldquoRiskReturn and Equilibrium Empirical TestsrdquoJournal of Political Economy 1973 81(3)pp 607ndash36

Fredman Albert J and Russ Wiles How mutualfunds work 2nd Edition New York NewYork Institute of Finance 1998

Gompers Paul A and Metrick Andrew ldquoInsti-tutional Investors and Equity Pricesrdquo Quar-terly Journal of Economics 2001 116(1) pp229ndash59

Grinblatt Mark and Titman Sheridan ldquoMu-tual Fund Performance An Analysis ofQuarterly Portfolio Holdingsrdquo Journal ofBusiness 1989 62(3) pp 393ndash 416

Grinblatt Mark Titman Sheridan and Werm-ers Russ ldquoMomentum Investment StrategiesPortfolio Performance and Herding A Studyof Mutual Fund Behaviorrdquo American Eco-nomic Review 1995 85(5) pp 1088ndash105

Grossman Sanford J and Hart Oliver D ldquoTheCosts and Benefits of Ownership A Theoryof Vertical and Lateral Integrationrdquo Journalof Political Economy 1986 94(4) pp 691ndash719

Gruber Martin J ldquoAnother Puzzle The Growthin Actively Managed Mutual Fundsrdquo Journalof Finance 1996 51(3) pp 783ndash810

Hart Oliver Firms contracts and financialstructure Clarendon Lectures in EconomicsOxford and New York Oxford UniversityPress Clarendon Press 1995

Hart Oliver and Moore John ldquoProperty Rightsand the Nature of the Firmrdquo Journal of

Political Economy 1990 98(6) pp 1119ndash58

Hechinger John ldquoDeciphering Fundsrsquo HiddenCostsrdquo Wall Street Journal New YorkMarch 17 2004 D1ndashD2

Holmstrom Bengt and Roberts John ldquoTheBoundaries of the Firm Revisitedrdquo Journalof Economic Perspectives 1998 12(4) pp73ndash94

Hong Harrison Kubik Jeffrey D and SolomonAmit ldquoSecurity Analystsrsquo Career Concernsand Herding of Earnings Forecastsrdquo RANDJournal of Economics 2000 31(1) pp 121ndash44

Hong Harrison Lim Terence and Stein JeremyC ldquoBad News Travels Slowly Size AnalystCoverage and the Profitability of MomentumStrategiesrdquo Journal of Finance 2000 55(1)pp 265ndash95

Investment Company Institute Mutual fund factbook Washington DC Investment CompanyInstitute 2000

IvKovic Zoran ldquoSpillovers in Mutual FundFamilies Is Blood Thicker Than WaterrdquoUnpublished Paper 2002

Jegadeesh Narasimhan and Titman SheridanldquoReturns to Buying Winners and Selling Los-ers Implications for Stock Market Effi-ciencyrdquo Journal of Finance 1993 48(1) pp65ndash91

Jensen Michael C ldquoThe Performance of MutualFunds in the Period 1945ndash1964rdquo Journal ofFinance 1968 50 pp 549ndash72

Lowenstein Roger ldquoFrightened Funds Is Therea Master in the Houserdquo Wall Street JournalNovember 20 1997 p C1

Lynch Anthony W and Musto David KldquoHow Investors Interpret Past Fund ReturnsrdquoJournal of Finance 2003 58(5) pp 2033ndash58

Malkiel Burton G ldquoReturns from Investing inEquity Mutual Funds 1971 to 1991rdquo Journalof Finance 1995 50(2) pp 549ndash72

Milgrom Paul and Roberts John ldquoAn EconomicApproach to Influence Activities in Organi-zationsrdquo American Journal of Sociology vol94 supplement 1988 pp 154ndash79

Newey Whitney K and West Kenneth D ldquoASimple Positive Semi-Definite Heteroske-dasticity and Autocorrelation ConsistentCovariance Matrixrdquo Econometrica 198755(3) pp 703ndash08


Perold Andre and Salomon Robert S ldquoTheRight Amount of Assets under Manage-mentrdquo Financial Analysts Journal 199147(3) pp 31ndash39

Pozen Robert C The mutual fund businessCambridge and London MIT Press 1998

Prather Larry J and Middleton Karen L ldquoAreN 1 Heads Better Than One The Case ofMutual Fund Managersrdquo Journal of Eco-nomic Behavior and Organization 200247(1) pp 103ndash20

Sharpe William F ldquoCapital Asset Prices ATheory of Market Equilibrium under Condi-tions of Riskrdquo Journal of Finance 196419(3) pp 425ndash42

Simester Duncan and Wernefelt Birger ldquoDe-terminants of Asset Ownership A Studyof Carpentry Traderdquo Unpublished Paper2000

Sirri Erik R and Tufano Peter ldquoCostly Search

and Mutual Fund Flowsrdquo Journal of Fi-nance 1998 53(5) pp 1589ndash622

Stein Jeremy C ldquoInformation Production andCapital Allocation Decentralized versus Hi-erarchical Firmsrdquo Journal of Finance 200257(5) pp 1891ndash921

Wermers Russ ldquoMutual Fund Performance AnEmpirical Decomposition into Stock-PickingTalent Style Transactions Costs and Ex-pensesrdquo Journal of Finance 2000 55(4) pp1655ndash95

Williamson Oliver E ldquoCorporate Finance andCorporate Governancerdquo Journal of Finance1988 43(3) pp 567ndash91

Williamson Oliver E Markets and hierarchiesAnalysis and antitrust implications NewYork Simon and Schuster 1975

Zheng Lu ldquoIs Money Smart A Study of Mu-tual Fund Investorsrsquo Fund Selection AbilityrdquoJournal of Finance 1999 54(3) pp 901ndash33


point out that there are advantages to scale suchas more resources for research and lower ex-pense ratios Others believe however that alarge asset base erodes fund performance be-cause of trading costs associated with liquidityor price impact (see eg Andre Perold andRobert S Salomon 1991 Roger Lowenstein1997) Whereas a small fund can easily put allof its money in its best ideas a lack of liquidityforces a large fund to have to invest in itsnot-so-good ideas and take larger positions perstock than is optimal thereby eroding perfor-mance Using a small sample of funds from1974 to 1984 Grinblatt and Sheridan Titman(1989) find mixed evidence that fund returnsdecline with fund size Needless to say there isno consensus on this issue

Using mutual fund data from 1962 to 1999we begin our investigation by using cross-sectional variation to see whether performancedepends on lagged fund size Since funds mayhave different styles we adjust for such heter-ogeneity by utilizing various performancebenchmarks that account for the possibility thatthey load differently on small cap stock valuestock and price momentum strategies More-over fund size may be correlated with suchother fund characteristics as fund age or turn-over and it may be these characteristics that aredriving performance Hence we regress the var-ious adjusted returns on lagged fund size (asmeasured by the log of total net assets undermanagement) and also include in the regres-sions a host of other observable fund character-istics including turnover age expense ratiototal load past-year fund inflows and past-yearreturns A number of studies warn that the re-ported returns of the smallest funds (those withless than $15 million in assets under manage-ment) might be upward biased We excludethese funds from our baseline sample in esti-mating these regressions

The regressions indicate that a fundrsquos perfor-mance is inversely correlated with its laggedassets under management For instance usingmonthly gross returns (before fees and expensesare deducted) a two-standard deviation shockin the log of a fundrsquos total assets under man-agement this month yields anywhere from a 54-to 77-basis-point movement in next monthrsquosfund return depending on the performancebenchmark (or about 65 to 96 basis points an-

nually) The corresponding figures for net fundreturns (after fees and expenses are deducted)are only slightly smaller To put these magni-tudes into some perspective the funds in oursample on average underperform the marketportfolio by about 96 basis points after fees andexpenses From this perspective a 65- to 96-basis-point annual spread in performance is notonly statistically significant but also economi-cally important1

Even after utilizing various performancebenchmarks and controlling for other observ-able fund characteristics there are still a num-ber of potential explanations that might beconsistent with the inverse relationship betweenscale and fund returns To further narrow the setof explanations we proceed to test a directimplication of the hypothesis that fund sizeerodes performance because of trading costsassociated with liquidity and price impact If theldquoliquidity hypothesisrdquo is true then size ought toerode performance much more for funds thathave to invest in small stocks which tend to beilliquid Consistent with this hypothesis we findthat fund size matters much more for the returnsamong such funds identified as ldquosmall caprdquofunds in our database than other funds2 Indeedfor other funds size does not significantly affectperformance This finding strongly indicatesthat liquidity plays an important role in thedocumented diseconomies of scale

We then delve deeper into the liquidity hy-pothesis by observing that liquidity means thatbig funds need to find more stock ideas thansmall funds do but liquidity itself may notcompletely explain why they cannot go aboutdoing this ie why they cannot scale Presum-ably a large fund can afford to hire additionalmanagers to cover more stocks It can therebygenerate additional good ideas so that it can take

1 As we describe below some theories suggest that thesmallest funds may have inferior performance to medium-sized ones because they are being run at a suboptimallysmall scale Because it is difficult to make inferences re-garding the performance of the smallest funds we do notattempt to measure such nonlinearities here

2 Throughout the paper we will sometimes refer to fundswith few assets under management as ldquosmall fundsrdquo andfunds that by virtue of their fund style have to invest insmall stocks as ldquosmall cap fundsrdquo So small cap funds arenot necessarily small funds Indeed most are actually quitelarge in terms of assets under management



















I Data


















($ million) [32] [72] [166] [548] [17971] [9258] [10227]LOGTNA 109 294 396 494 645 387 457



( per year) [8303] [7400] [6860] [6428] [5456] [7184] [6721]AGE 817 1190 1482 1843 2516 1567 1757

(years) [854] [1073] [1285] [1438] [1506] [1396] [1433]EXPRATIO 129 108 094 085 068 097 089

( per year) [111] [059] [046] [037] [031] [068] [048]TOTLOAD 341 419 432 457 528 436 459

() [332] [332] [334] [339] [288] [336] [332]FLOW 3079 3066 2697 2127 1354 2467 2313

( per year) [11355] [11336] [10166] [8408] [5904] [10264] [9660]










FUNDRET 009 002 003 006 006 001 002(Gross) [304] [264] [261] [246] [200] [262] [248]

FUNDRET 002 007 005 013 012 008 009(Net) [304] [264] [261] [246] [200] [262] [248]









II Methodology











(1)








i3 HMLt it t 1 T



t 1 T




FactorMeanreturn

SD ofreturn

Cross-correlations

VWRF SMB HML MOM12

VWRF 058 437 100 032 039 002SMB 017 290 100 016 030HML 034 263 100 015MOM12 096 388 100


Portfolio

CAPM

Alpha VWRF

1 (small) 004 0872 002 0913 001 0934 009 0925 (large) 007 091


Portfolio

3-Factor model

HML

4-Factor model


1 (small) 001 082 029 003 002 082 030 004 0032 003 084 027 001 009 085 030 000 0053 001 087 022 005 006 087 025 003 0064 006 087 018 005 013 087 020 004 0065 (large) 005 088 008 006 010 088 010 005 005








Xit 1 it i 1 N






Xit 1 it i 1 N


III Results











INTERCEPT 0056 0087 0080 0019 0026 0056 0049 0011(096) (163) (128) (030) (044) (105) (079) (018)

LOGTNAit1 0028 0028 0023 0020 0025 0025 0020 0018(302) (304) (242) (215) (275) (276) (216) (189)

LOGFAMSIZEit1 0007 0007 0007 0007 0007 0007 0007 0007(226) (227) (222) (221) (233) (234) (229) (228)

TURNOVERit1 0000 0000 0000 0000 0000 0000 0000 0000(085) (086) (085) (083) (077) (078) (077) (075)

AGEit1 0001 0001 0001 0001 0000 0000 0001 0001(062) (062) (064) (063) (052) (053) (055) (054)

EXPRATIOit1 0004 0004 0007 0007 0039 0038 0041 0041(011) (009) (018) (018) (097) (095) (104) (105)

TOTLOADit1 0003 0003 0003 0003 0003 0003 0003 0003(126) (125) (126) (129) (121) (120) (121) (125)

FLOWit1 0000 0000 0000 0000 0000 0000 0000 0000(050) (050) (051) (049) (047) (047) (048) (046)

LAGFUNDRETit1 0029 0028 0028 0029 0029 0029 0029 0029(600) (598) (600) (599) (603) (601) (603) (602)

No of months 444 444 444 444 444 444 444 444



















LOGTNAit1 0023 0026 0019 0021 0020 0022 0015 0017(259) (294) (222) (248) (216) (249) (176) (202)

LOGFAMSIZEit1 0006 0006 0006 0006 0006 0006 0006 0006(199) (203) (199) (205) (210) (215) (211) (217)

No of months 444 444 444 444 444 444 444 444




LOGTNAit1 0436 0440 0345 0312 0402 0406 0312 0280(326) (327) (240) (219) (307) (307) (220) (199)

LOGFAMSIZEit1 0088 0089 0088 0088 0090 0091 0090 0090(205) (206) (205) (207) (210) (212) (211) (212)

No of years 37 37 37 37 37 37 37 37




LOGTNAit1 0035 0034 0021 0019 0032 0032 0019 0016(241) (242) (151) (133) (222) (223) (132) (115)

LOGFAMSIZEit1 0002 0002 0002 0002 0002 0002 0002 0002(049) (050) (045) (044) (056) (058) (053) (052)

No of months 216 216 216 216 216 216 216 216




LOGTNAit1 0021 0022 0024 0022 0019 0019 0022 0019(184) (186) (193) (171) (164) (165) (174) (153)

LOGFAMSIZEit1 0012 0012 0012 0012 0012 0012 0012 0012(253) (253) (250) (250) (255) (255) (252) (252)

No of months 228 228 228 228 228 228 228 228





Portfolio





















INTERCEPT 0205 0257 0320 0263 0175 0226 0289 0231(127) (153) (244) (198) (108) (134) (220) (175)

INDnot SCG 0250 0252 0260 0262 0249 0251 0259 0261(154) (155) (162) (163) (153) (154) (161) (162)

LOGTNAit1 0058 0058 0063 0061 0055 0056 0060 0058(315) (317) (340) (329) (303) (305) (328) (317)


LOGFAMSIZEit1 0012 0012 0011 0011 0012 0012 0012 0012(257) (257) (254) (254) (259) (259) (257) (256)

TURNOVERit1 0000 0000 0000 0000 0000 0000 0000 0000(162) (161) (162) (161) (157) (156) (157) (157)

AGEit1 0001 0001 0001 0001 0001 0001 0001 0001(098) (097) (098) (097) (087) (087) (088) (087)

EXPRATIOit1 0018 0018 0019 0019 0062 0062 0064 0063(044) (043) (046) (044) (149) (148) (152) (151)

TOTLOADit1 0001 0001 0001 0001 0001 0001 0001 0001(054) (054) (052) (052) (048) (047) (046) (046)

FLOWit1 0000 0000 0000 0000 0000 0000 0000 0000(096) (095) (090) (092) (102) (102) (096) (099)

FUNDRETit1 0022 0022 0021 0021 0022 0022 0022 0022(345) (345) (343) (342) (348) (348) (345) (345)

No of months 228 228 228 228 228 228 228 228




INTERCEPT 0016 0068 0128 0068 0017 0035 0095 0034(019) (084) (172) (092) (020) (043) (127) (047)

INDLC 0119 0121 0119 0120 0112 0114 0112 0113(131) (133) (133) (134) (123) (125) (124) (126)

LOGTNAit1 0029 0029 0032 0029 0027 0027 0029 0027(222) (224) (245) (225) (202) (205) (226) (205)

LOGTNAit1 0031 0031 0031 0031 0030 0030 0030 0030INDLC (229) (232) (234) (236) (223) (226) (228) (230)

LOGFAMSIZEit1 0011 0011 0011 0011 0011 0011 0011 0011(254) (254) (251) (251) (258) (258) (255) (255)

TURNOVERit1 0000 0000 0000 0000 0000 0000 0000 0000(180) (179) (180) (179) (175) (175) (176) (175)

AGEit1 0001 0001 0001 0001 0001 0001 0001 0001(126) (125) (126) (125) (117) (117) (117) (116)

EXPRATIOit1 0020 0019 0021 0020 0065 0065 0066 0065(049) (048) (052) (050) (163) (162) (166) (164)

TOTLOADit1 0001 0001 0001 0001 0001 0001 0001 0001(035) (035) (034) (035) (030) (030) (029) (029)

FLOWit1 0000 0000 0000 0000 0000 0000 0000 0000(084) (082) (077) (079) (089) (088) (083) (085)

FUNDRETit1 0020 0020 0020 0020 0021 0020 0020 0020(310) (309) (307) (307) (313) (312) (310) (309)

No of months 228 228 228 228 228 228 228 228












LOGTNAit1 0051 0052 0056 0054 0048 0049 0053 0051(283) (286) (313) (302) (270) (272) (301) (289)


LOGFAMSIZEit1 0007 0007 0007 0007 0007 0007 0007 0007(070) (070) (069) (070) (070) (070) (069) (070)


No of months 228 228 228 228 228 228 228 228




LOGTNAit1 0067 0070 0071 0072 0064 0067 0068 0069(367) (383) (387) (391) (351) (367) (371) (376)


LOGFAMSIZEit1 0012 0012 0012 0012 0012 0012 0012 0012(113) (113) (111) (111) (115) (115) (113) (113)


No of months 228 228 228 228 228 228 228 228




































(1) (2) (3)

SMALLFUND 00759 00580 00584(949) (674) (679)

SMALLCAPFUND 00515 00340 00341(628) (386) (388)














(1) (2) (3)

SMALLFUND 00258 00143 00150(183) (094) (099)

SMALLCAPFUND 01190 01078 01079(821) (695) (696)














STOCKS


MULTI-MANAGER 00102 00190(175) (173)


















V Conclusion





INTERCEPT 0069 0018 0039 0008 0092 0042 0014 0032(066) (019) (025) (005) (088) (044) (009) (021)

MULTI-MANAGERi 0040 0040 0041 0041 0040 0041 0042 0042(222) (223) (230) (230) (226) (227) (234) (234)

LOGTNAit1 0024 0023 0026 0025 0022 0022 0025 0023(234) (230) (198) (183) (219) (215) (186) (170)

LOGFAMSIZEit1 0016 0016 0016 0016 0016 0016 0016 0016(238) (238) (238) (238) (240) (239) (240) (239)

TURNOVERit1 0000 0000 0000 0000 0000 0000 0000 0000(128) (129) (128) (129) (126) (126) (126) (126)

AGEit1 0001 0001 0001 0001 0001 0001 0001 0001(064) (065) (064) (064) (060) (061) (059) (059)

EXPRATIOit1 0002 0002 0003 0003 0043 0043 0045 0045(003) (003) (004) (004) (054) (054) (055) (055)

TOTLOADit1 0002 0002 0002 0002 0002 0002 0002 0002(028) (029) (030) (029) (036) (036) (037) (037)

FLOWit1 0000 0000 0000 0000 0000 0000 0000 0000(164) (168) (168) (169) (164) (168) (168) (169)

FUNDRETit1 0030 0030 0029 0029 0031 0030 0030 0030(338) (338) (334) (334) (341) (341) (337) (337)

No of months 96 96 96 96 96 96 96 96





REFERENCES













































































I Data


















($ million) [32] [72] [166] [548] [17971] [9258] [10227]LOGTNA 109 294 396 494 645 387 457



( per year) [8303] [7400] [6860] [6428] [5456] [7184] [6721]AGE 817 1190 1482 1843 2516 1567 1757

(years) [854] [1073] [1285] [1438] [1506] [1396] [1433]EXPRATIO 129 108 094 085 068 097 089

( per year) [111] [059] [046] [037] [031] [068] [048]TOTLOAD 341 419 432 457 528 436 459

() [332] [332] [334] [339] [288] [336] [332]FLOW 3079 3066 2697 2127 1354 2467 2313

( per year) [11355] [11336] [10166] [8408] [5904] [10264] [9660]










FUNDRET 009 002 003 006 006 001 002(Gross) [304] [264] [261] [246] [200] [262] [248]

FUNDRET 002 007 005 013 012 008 009(Net) [304] [264] [261] [246] [200] [262] [248]









II Methodology











(1)








i3 HMLt it t 1 T



t 1 T




FactorMeanreturn

SD ofreturn

Cross-correlations

VWRF SMB HML MOM12

VWRF 058 437 100 032 039 002SMB 017 290 100 016 030HML 034 263 100 015MOM12 096 388 100


Portfolio

CAPM

Alpha VWRF

1 (small) 004 0872 002 0913 001 0934 009 0925 (large) 007 091


Portfolio

3-Factor model

HML

4-Factor model


1 (small) 001 082 029 003 002 082 030 004 0032 003 084 027 001 009 085 030 000 0053 001 087 022 005 006 087 025 003 0064 006 087 018 005 013 087 020 004 0065 (large) 005 088 008 006 010 088 010 005 005








Xit 1 it i 1 N






Xit 1 it i 1 N


III Results











INTERCEPT 0056 0087 0080 0019 0026 0056 0049 0011(096) (163) (128) (030) (044) (105) (079) (018)

LOGTNAit1 0028 0028 0023 0020 0025 0025 0020 0018(302) (304) (242) (215) (275) (276) (216) (189)

LOGFAMSIZEit1 0007 0007 0007 0007 0007 0007 0007 0007(226) (227) (222) (221) (233) (234) (229) (228)

TURNOVERit1 0000 0000 0000 0000 0000 0000 0000 0000(085) (086) (085) (083) (077) (078) (077) (075)

AGEit1 0001 0001 0001 0001 0000 0000 0001 0001(062) (062) (064) (063) (052) (053) (055) (054)

EXPRATIOit1 0004 0004 0007 0007 0039 0038 0041 0041(011) (009) (018) (018) (097) (095) (104) (105)

TOTLOADit1 0003 0003 0003 0003 0003 0003 0003 0003(126) (125) (126) (129) (121) (120) (121) (125)

FLOWit1 0000 0000 0000 0000 0000 0000 0000 0000(050) (050) (051) (049) (047) (047) (048) (046)

LAGFUNDRETit1 0029 0028 0028 0029 0029 0029 0029 0029(600) (598) (600) (599) (603) (601) (603) (602)

No of months 444 444 444 444 444 444 444 444



















LOGTNAit1 0023 0026 0019 0021 0020 0022 0015 0017(259) (294) (222) (248) (216) (249) (176) (202)

LOGFAMSIZEit1 0006 0006 0006 0006 0006 0006 0006 0006(199) (203) (199) (205) (210) (215) (211) (217)

No of months 444 444 444 444 444 444 444 444




LOGTNAit1 0436 0440 0345 0312 0402 0406 0312 0280(326) (327) (240) (219) (307) (307) (220) (199)

LOGFAMSIZEit1 0088 0089 0088 0088 0090 0091 0090 0090(205) (206) (205) (207) (210) (212) (211) (212)

No of years 37 37 37 37 37 37 37 37




LOGTNAit1 0035 0034 0021 0019 0032 0032 0019 0016(241) (242) (151) (133) (222) (223) (132) (115)

LOGFAMSIZEit1 0002 0002 0002 0002 0002 0002 0002 0002(049) (050) (045) (044) (056) (058) (053) (052)

No of months 216 216 216 216 216 216 216 216




LOGTNAit1 0021 0022 0024 0022 0019 0019 0022 0019(184) (186) (193) (171) (164) (165) (174) (153)

LOGFAMSIZEit1 0012 0012 0012 0012 0012 0012 0012 0012(253) (253) (250) (250) (255) (255) (252) (252)

No of months 228 228 228 228 228 228 228 228





Portfolio





















INTERCEPT 0205 0257 0320 0263 0175 0226 0289 0231(127) (153) (244) (198) (108) (134) (220) (175)

INDnot SCG 0250 0252 0260 0262 0249 0251 0259 0261(154) (155) (162) (163) (153) (154) (161) (162)

LOGTNAit1 0058 0058 0063 0061 0055 0056 0060 0058(315) (317) (340) (329) (303) (305) (328) (317)


LOGFAMSIZEit1 0012 0012 0011 0011 0012 0012 0012 0012(257) (257) (254) (254) (259) (259) (257) (256)

TURNOVERit1 0000 0000 0000 0000 0000 0000 0000 0000(162) (161) (162) (161) (157) (156) (157) (157)

AGEit1 0001 0001 0001 0001 0001 0001 0001 0001(098) (097) (098) (097) (087) (087) (088) (087)

EXPRATIOit1 0018 0018 0019 0019 0062 0062 0064 0063(044) (043) (046) (044) (149) (148) (152) (151)

TOTLOADit1 0001 0001 0001 0001 0001 0001 0001 0001(054) (054) (052) (052) (048) (047) (046) (046)

FLOWit1 0000 0000 0000 0000 0000 0000 0000 0000(096) (095) (090) (092) (102) (102) (096) (099)

FUNDRETit1 0022 0022 0021 0021 0022 0022 0022 0022(345) (345) (343) (342) (348) (348) (345) (345)

No of months 228 228 228 228 228 228 228 228




INTERCEPT 0016 0068 0128 0068 0017 0035 0095 0034(019) (084) (172) (092) (020) (043) (127) (047)

INDLC 0119 0121 0119 0120 0112 0114 0112 0113(131) (133) (133) (134) (123) (125) (124) (126)

LOGTNAit1 0029 0029 0032 0029 0027 0027 0029 0027(222) (224) (245) (225) (202) (205) (226) (205)

LOGTNAit1 0031 0031 0031 0031 0030 0030 0030 0030INDLC (229) (232) (234) (236) (223) (226) (228) (230)

LOGFAMSIZEit1 0011 0011 0011 0011 0011 0011 0011 0011(254) (254) (251) (251) (258) (258) (255) (255)

TURNOVERit1 0000 0000 0000 0000 0000 0000 0000 0000(180) (179) (180) (179) (175) (175) (176) (175)

AGEit1 0001 0001 0001 0001 0001 0001 0001 0001(126) (125) (126) (125) (117) (117) (117) (116)

EXPRATIOit1 0020 0019 0021 0020 0065 0065 0066 0065(049) (048) (052) (050) (163) (162) (166) (164)

TOTLOADit1 0001 0001 0001 0001 0001 0001 0001 0001(035) (035) (034) (035) (030) (030) (029) (029)

FLOWit1 0000 0000 0000 0000 0000 0000 0000 0000(084) (082) (077) (079) (089) (088) (083) (085)

FUNDRETit1 0020 0020 0020 0020 0021 0020 0020 0020(310) (309) (307) (307) (313) (312) (310) (309)

No of months 228 228 228 228 228 228 228 228












LOGTNAit1 0051 0052 0056 0054 0048 0049 0053 0051(283) (286) (313) (302) (270) (272) (301) (289)


LOGFAMSIZEit1 0007 0007 0007 0007 0007 0007 0007 0007(070) (070) (069) (070) (070) (070) (069) (070)


No of months 228 228 228 228 228 228 228 228




LOGTNAit1 0067 0070 0071 0072 0064 0067 0068 0069(367) (383) (387) (391) (351) (367) (371) (376)


LOGFAMSIZEit1 0012 0012 0012 0012 0012 0012 0012 0012(113) (113) (111) (111) (115) (115) (113) (113)


No of months 228 228 228 228 228 228 228 228




































(1) (2) (3)

SMALLFUND 00759 00580 00584(949) (674) (679)

SMALLCAPFUND 00515 00340 00341(628) (386) (388)














(1) (2) (3)

SMALLFUND 00258 00143 00150(183) (094) (099)

SMALLCAPFUND 01190 01078 01079(821) (695) (696)














STOCKS


MULTI-MANAGER 00102 00190(175) (173)


















V Conclusion





INTERCEPT 0069 0018 0039 0008 0092 0042 0014 0032(066) (019) (025) (005) (088) (044) (009) (021)

MULTI-MANAGERi 0040 0040 0041 0041 0040 0041 0042 0042(222) (223) (230) (230) (226) (227) (234) (234)

LOGTNAit1 0024 0023 0026 0025 0022 0022 0025 0023(234) (230) (198) (183) (219) (215) (186) (170)

LOGFAMSIZEit1 0016 0016 0016 0016 0016 0016 0016 0016(238) (238) (238) (238) (240) (239) (240) (239)

TURNOVERit1 0000 0000 0000 0000 0000 0000 0000 0000(128) (129) (128) (129) (126) (126) (126) (126)

AGEit1 0001 0001 0001 0001 0001 0001 0001 0001(064) (065) (064) (064) (060) (061) (059) (059)

EXPRATIOit1 0002 0002 0003 0003 0043 0043 0045 0045(003) (003) (004) (004) (054) (054) (055) (055)

TOTLOADit1 0002 0002 0002 0002 0002 0002 0002 0002(028) (029) (030) (029) (036) (036) (037) (037)

FLOWit1 0000 0000 0000 0000 0000 0000 0000 0000(164) (168) (168) (169) (164) (168) (168) (169)

FUNDRETit1 0030 0030 0029 0029 0031 0030 0030 0030(338) (338) (334) (334) (341) (341) (337) (337)

No of months 96 96 96 96 96 96 96 96





REFERENCES










































































($ million) [32] [72] [166] [548] [17971] [9258] [10227]LOGTNA 109 294 396 494 645 387 457



( per year) [8303] [7400] [6860] [6428] [5456] [7184] [6721]AGE 817 1190 1482 1843 2516 1567 1757

(years) [854] [1073] [1285] [1438] [1506] [1396] [1433]EXPRATIO 129 108 094 085 068 097 089

( per year) [111] [059] [046] [037] [031] [068] [048]TOTLOAD 341 419 432 457 528 436 459

() [332] [332] [334] [339] [288] [336] [332]FLOW 3079 3066 2697 2127 1354 2467 2313

( per year) [11355] [11336] [10166] [8408] [5904] [10264] [9660]










FUNDRET 009 002 003 006 006 001 002(Gross) [304] [264] [261] [246] [200] [262] [248]

FUNDRET 002 007 005 013 012 008 009(Net) [304] [264] [261] [246] [200] [262] [248]









II Methodology











(1)








i3 HMLt it t 1 T



t 1 T




FactorMeanreturn

SD ofreturn

Cross-correlations

VWRF SMB HML MOM12

VWRF 058 437 100 032 039 002SMB 017 290 100 016 030HML 034 263 100 015MOM12 096 388 100


Portfolio

CAPM

Alpha VWRF

1 (small) 004 0872 002 0913 001 0934 009 0925 (large) 007 091


Portfolio

3-Factor model

HML

4-Factor model


1 (small) 001 082 029 003 002 082 030 004 0032 003 084 027 001 009 085 030 000 0053 001 087 022 005 006 087 025 003 0064 006 087 018 005 013 087 020 004 0065 (large) 005 088 008 006 010 088 010 005 005








Xit 1 it i 1 N






Xit 1 it i 1 N


III Results











INTERCEPT 0056 0087 0080 0019 0026 0056 0049 0011(096) (163) (128) (030) (044) (105) (079) (018)

LOGTNAit1 0028 0028 0023 0020 0025 0025 0020 0018(302) (304) (242) (215) (275) (276) (216) (189)

LOGFAMSIZEit1 0007 0007 0007 0007 0007 0007 0007 0007(226) (227) (222) (221) (233) (234) (229) (228)

TURNOVERit1 0000 0000 0000 0000 0000 0000 0000 0000(085) (086) (085) (083) (077) (078) (077) (075)

AGEit1 0001 0001 0001 0001 0000 0000 0001 0001(062) (062) (064) (063) (052) (053) (055) (054)

EXPRATIOit1 0004 0004 0007 0007 0039 0038 0041 0041(011) (009) (018) (018) (097) (095) (104) (105)

TOTLOADit1 0003 0003 0003 0003 0003 0003 0003 0003(126) (125) (126) (129) (121) (120) (121) (125)

FLOWit1 0000 0000 0000 0000 0000 0000 0000 0000(050) (050) (051) (049) (047) (047) (048) (046)

LAGFUNDRETit1 0029 0028 0028 0029 0029 0029 0029 0029(600) (598) (600) (599) (603) (601) (603) (602)

No of months 444 444 444 444 444 444 444 444



















LOGTNAit1 0023 0026 0019 0021 0020 0022 0015 0017(259) (294) (222) (248) (216) (249) (176) (202)

LOGFAMSIZEit1 0006 0006 0006 0006 0006 0006 0006 0006(199) (203) (199) (205) (210) (215) (211) (217)

No of months 444 444 444 444 444 444 444 444




LOGTNAit1 0436 0440 0345 0312 0402 0406 0312 0280(326) (327) (240) (219) (307) (307) (220) (199)

LOGFAMSIZEit1 0088 0089 0088 0088 0090 0091 0090 0090(205) (206) (205) (207) (210) (212) (211) (212)

No of years 37 37 37 37 37 37 37 37




LOGTNAit1 0035 0034 0021 0019 0032 0032 0019 0016(241) (242) (151) (133) (222) (223) (132) (115)

LOGFAMSIZEit1 0002 0002 0002 0002 0002 0002 0002 0002(049) (050) (045) (044) (056) (058) (053) (052)

No of months 216 216 216 216 216 216 216 216




LOGTNAit1 0021 0022 0024 0022 0019 0019 0022 0019(184) (186) (193) (171) (164) (165) (174) (153)

LOGFAMSIZEit1 0012 0012 0012 0012 0012 0012 0012 0012(253) (253) (250) (250) (255) (255) (252) (252)

No of months 228 228 228 228 228 228 228 228





Portfolio





















INTERCEPT 0205 0257 0320 0263 0175 0226 0289 0231(127) (153) (244) (198) (108) (134) (220) (175)

INDnot SCG 0250 0252 0260 0262 0249 0251 0259 0261(154) (155) (162) (163) (153) (154) (161) (162)

LOGTNAit1 0058 0058 0063 0061 0055 0056 0060 0058(315) (317) (340) (329) (303) (305) (328) (317)


LOGFAMSIZEit1 0012 0012 0011 0011 0012 0012 0012 0012(257) (257) (254) (254) (259) (259) (257) (256)

TURNOVERit1 0000 0000 0000 0000 0000 0000 0000 0000(162) (161) (162) (161) (157) (156) (157) (157)

AGEit1 0001 0001 0001 0001 0001 0001 0001 0001(098) (097) (098) (097) (087) (087) (088) (087)

EXPRATIOit1 0018 0018 0019 0019 0062 0062 0064 0063(044) (043) (046) (044) (149) (148) (152) (151)

TOTLOADit1 0001 0001 0001 0001 0001 0001 0001 0001(054) (054) (052) (052) (048) (047) (046) (046)

FLOWit1 0000 0000 0000 0000 0000 0000 0000 0000(096) (095) (090) (092) (102) (102) (096) (099)

FUNDRETit1 0022 0022 0021 0021 0022 0022 0022 0022(345) (345) (343) (342) (348) (348) (345) (345)

No of months 228 228 228 228 228 228 228 228




INTERCEPT 0016 0068 0128 0068 0017 0035 0095 0034(019) (084) (172) (092) (020) (043) (127) (047)

INDLC 0119 0121 0119 0120 0112 0114 0112 0113(131) (133) (133) (134) (123) (125) (124) (126)

LOGTNAit1 0029 0029 0032 0029 0027 0027 0029 0027(222) (224) (245) (225) (202) (205) (226) (205)

LOGTNAit1 0031 0031 0031 0031 0030 0030 0030 0030INDLC (229) (232) (234) (236) (223) (226) (228) (230)

LOGFAMSIZEit1 0011 0011 0011 0011 0011 0011 0011 0011(254) (254) (251) (251) (258) (258) (255) (255)

TURNOVERit1 0000 0000 0000 0000 0000 0000 0000 0000(180) (179) (180) (179) (175) (175) (176) (175)

AGEit1 0001 0001 0001 0001 0001 0001 0001 0001(126) (125) (126) (125) (117) (117) (117) (116)

EXPRATIOit1 0020 0019 0021 0020 0065 0065 0066 0065(049) (048) (052) (050) (163) (162) (166) (164)

TOTLOADit1 0001 0001 0001 0001 0001 0001 0001 0001(035) (035) (034) (035) (030) (030) (029) (029)

FLOWit1 0000 0000 0000 0000 0000 0000 0000 0000(084) (082) (077) (079) (089) (088) (083) (085)

FUNDRETit1 0020 0020 0020 0020 0021 0020 0020 0020(310) (309) (307) (307) (313) (312) (310) (309)

No of months 228 228 228 228 228 228 228 228












LOGTNAit1 0051 0052 0056 0054 0048 0049 0053 0051(283) (286) (313) (302) (270) (272) (301) (289)


LOGFAMSIZEit1 0007 0007 0007 0007 0007 0007 0007 0007(070) (070) (069) (070) (070) (070) (069) (070)


No of months 228 228 228 228 228 228 228 228




LOGTNAit1 0067 0070 0071 0072 0064 0067 0068 0069(367) (383) (387) (391) (351) (367) (371) (376)


LOGFAMSIZEit1 0012 0012 0012 0012 0012 0012 0012 0012(113) (113) (111) (111) (115) (115) (113) (113)


No of months 228 228 228 228 228 228 228 228




































(1) (2) (3)

SMALLFUND 00759 00580 00584(949) (674) (679)

SMALLCAPFUND 00515 00340 00341(628) (386) (388)














(1) (2) (3)

SMALLFUND 00258 00143 00150(183) (094) (099)

SMALLCAPFUND 01190 01078 01079(821) (695) (696)














STOCKS


MULTI-MANAGER 00102 00190(175) (173)


















V Conclusion





INTERCEPT 0069 0018 0039 0008 0092 0042 0014 0032(066) (019) (025) (005) (088) (044) (009) (021)

MULTI-MANAGERi 0040 0040 0041 0041 0040 0041 0042 0042(222) (223) (230) (230) (226) (227) (234) (234)

LOGTNAit1 0024 0023 0026 0025 0022 0022 0025 0023(234) (230) (198) (183) (219) (215) (186) (170)

LOGFAMSIZEit1 0016 0016 0016 0016 0016 0016 0016 0016(238) (238) (238) (238) (240) (239) (240) (239)

TURNOVERit1 0000 0000 0000 0000 0000 0000 0000 0000(128) (129) (128) (129) (126) (126) (126) (126)

AGEit1 0001 0001 0001 0001 0001 0001 0001 0001(064) (065) (064) (064) (060) (061) (059) (059)

EXPRATIOit1 0002 0002 0003 0003 0043 0043 0045 0045(003) (003) (004) (004) (054) (054) (055) (055)

TOTLOADit1 0002 0002 0002 0002 0002 0002 0002 0002(028) (029) (030) (029) (036) (036) (037) (037)

FLOWit1 0000 0000 0000 0000 0000 0000 0000 0000(164) (168) (168) (169) (164) (168) (168) (169)

FUNDRETit1 0030 0030 0029 0029 0031 0030 0030 0030(338) (338) (334) (334) (341) (341) (337) (337)

No of months 96 96 96 96 96 96 96 96





REFERENCES


































































II Methodology











(1)








i3 HMLt it t 1 T



t 1 T




FactorMeanreturn

SD ofreturn

Cross-correlations

VWRF SMB HML MOM12

VWRF 058 437 100 032 039 002SMB 017 290 100 016 030HML 034 263 100 015MOM12 096 388 100


Portfolio

CAPM

Alpha VWRF

1 (small) 004 0872 002 0913 001 0934 009 0925 (large) 007 091


Portfolio

3-Factor model

HML

4-Factor model


1 (small) 001 082 029 003 002 082 030 004 0032 003 084 027 001 009 085 030 000 0053 001 087 022 005 006 087 025 003 0064 006 087 018 005 013 087 020 004 0065 (large) 005 088 008 006 010 088 010 005 005








Xit 1 it i 1 N






Xit 1 it i 1 N


III Results











INTERCEPT 0056 0087 0080 0019 0026 0056 0049 0011(096) (163) (128) (030) (044) (105) (079) (018)

LOGTNAit1 0028 0028 0023 0020 0025 0025 0020 0018(302) (304) (242) (215) (275) (276) (216) (189)

LOGFAMSIZEit1 0007 0007 0007 0007 0007 0007 0007 0007(226) (227) (222) (221) (233) (234) (229) (228)

TURNOVERit1 0000 0000 0000 0000 0000 0000 0000 0000(085) (086) (085) (083) (077) (078) (077) (075)

AGEit1 0001 0001 0001 0001 0000 0000 0001 0001(062) (062) (064) (063) (052) (053) (055) (054)

EXPRATIOit1 0004 0004 0007 0007 0039 0038 0041 0041(011) (009) (018) (018) (097) (095) (104) (105)

TOTLOADit1 0003 0003 0003 0003 0003 0003 0003 0003(126) (125) (126) (129) (121) (120) (121) (125)

FLOWit1 0000 0000 0000 0000 0000 0000 0000 0000(050) (050) (051) (049) (047) (047) (048) (046)

LAGFUNDRETit1 0029 0028 0028 0029 0029 0029 0029 0029(600) (598) (600) (599) (603) (601) (603) (602)

No of months 444 444 444 444 444 444 444 444



















LOGTNAit1 0023 0026 0019 0021 0020 0022 0015 0017(259) (294) (222) (248) (216) (249) (176) (202)

LOGFAMSIZEit1 0006 0006 0006 0006 0006 0006 0006 0006(199) (203) (199) (205) (210) (215) (211) (217)

No of months 444 444 444 444 444 444 444 444




LOGTNAit1 0436 0440 0345 0312 0402 0406 0312 0280(326) (327) (240) (219) (307) (307) (220) (199)

LOGFAMSIZEit1 0088 0089 0088 0088 0090 0091 0090 0090(205) (206) (205) (207) (210) (212) (211) (212)

No of years 37 37 37 37 37 37 37 37




LOGTNAit1 0035 0034 0021 0019 0032 0032 0019 0016(241) (242) (151) (133) (222) (223) (132) (115)

LOGFAMSIZEit1 0002 0002 0002 0002 0002 0002 0002 0002(049) (050) (045) (044) (056) (058) (053) (052)

No of months 216 216 216 216 216 216 216 216




LOGTNAit1 0021 0022 0024 0022 0019 0019 0022 0019(184) (186) (193) (171) (164) (165) (174) (153)

LOGFAMSIZEit1 0012 0012 0012 0012 0012 0012 0012 0012(253) (253) (250) (250) (255) (255) (252) (252)

No of months 228 228 228 228 228 228 228 228





Portfolio





















INTERCEPT 0205 0257 0320 0263 0175 0226 0289 0231(127) (153) (244) (198) (108) (134) (220) (175)

INDnot SCG 0250 0252 0260 0262 0249 0251 0259 0261(154) (155) (162) (163) (153) (154) (161) (162)

LOGTNAit1 0058 0058 0063 0061 0055 0056 0060 0058(315) (317) (340) (329) (303) (305) (328) (317)


LOGFAMSIZEit1 0012 0012 0011 0011 0012 0012 0012 0012(257) (257) (254) (254) (259) (259) (257) (256)

TURNOVERit1 0000 0000 0000 0000 0000 0000 0000 0000(162) (161) (162) (161) (157) (156) (157) (157)

AGEit1 0001 0001 0001 0001 0001 0001 0001 0001(098) (097) (098) (097) (087) (087) (088) (087)

EXPRATIOit1 0018 0018 0019 0019 0062 0062 0064 0063(044) (043) (046) (044) (149) (148) (152) (151)

TOTLOADit1 0001 0001 0001 0001 0001 0001 0001 0001(054) (054) (052) (052) (048) (047) (046) (046)

FLOWit1 0000 0000 0000 0000 0000 0000 0000 0000(096) (095) (090) (092) (102) (102) (096) (099)

FUNDRETit1 0022 0022 0021 0021 0022 0022 0022 0022(345) (345) (343) (342) (348) (348) (345) (345)

No of months 228 228 228 228 228 228 228 228




INTERCEPT 0016 0068 0128 0068 0017 0035 0095 0034(019) (084) (172) (092) (020) (043) (127) (047)

INDLC 0119 0121 0119 0120 0112 0114 0112 0113(131) (133) (133) (134) (123) (125) (124) (126)

LOGTNAit1 0029 0029 0032 0029 0027 0027 0029 0027(222) (224) (245) (225) (202) (205) (226) (205)

LOGTNAit1 0031 0031 0031 0031 0030 0030 0030 0030INDLC (229) (232) (234) (236) (223) (226) (228) (230)

LOGFAMSIZEit1 0011 0011 0011 0011 0011 0011 0011 0011(254) (254) (251) (251) (258) (258) (255) (255)

TURNOVERit1 0000 0000 0000 0000 0000 0000 0000 0000(180) (179) (180) (179) (175) (175) (176) (175)

AGEit1 0001 0001 0001 0001 0001 0001 0001 0001(126) (125) (126) (125) (117) (117) (117) (116)

EXPRATIOit1 0020 0019 0021 0020 0065 0065 0066 0065(049) (048) (052) (050) (163) (162) (166) (164)

TOTLOADit1 0001 0001 0001 0001 0001 0001 0001 0001(035) (035) (034) (035) (030) (030) (029) (029)

FLOWit1 0000 0000 0000 0000 0000 0000 0000 0000(084) (082) (077) (079) (089) (088) (083) (085)

FUNDRETit1 0020 0020 0020 0020 0021 0020 0020 0020(310) (309) (307) (307) (313) (312) (310) (309)

No of months 228 228 228 228 228 228 228 228












LOGTNAit1 0051 0052 0056 0054 0048 0049 0053 0051(283) (286) (313) (302) (270) (272) (301) (289)


LOGFAMSIZEit1 0007 0007 0007 0007 0007 0007 0007 0007(070) (070) (069) (070) (070) (070) (069) (070)


No of months 228 228 228 228 228 228 228 228




LOGTNAit1 0067 0070 0071 0072 0064 0067 0068 0069(367) (383) (387) (391) (351) (367) (371) (376)


LOGFAMSIZEit1 0012 0012 0012 0012 0012 0012 0012 0012(113) (113) (111) (111) (115) (115) (113) (113)


No of months 228 228 228 228 228 228 228 228




































(1) (2) (3)

SMALLFUND 00759 00580 00584(949) (674) (679)

SMALLCAPFUND 00515 00340 00341(628) (386) (388)














(1) (2) (3)

SMALLFUND 00258 00143 00150(183) (094) (099)

SMALLCAPFUND 01190 01078 01079(821) (695) (696)














STOCKS


MULTI-MANAGER 00102 00190(175) (173)


















V Conclusion





INTERCEPT 0069 0018 0039 0008 0092 0042 0014 0032(066) (019) (025) (005) (088) (044) (009) (021)

MULTI-MANAGERi 0040 0040 0041 0041 0040 0041 0042 0042(222) (223) (230) (230) (226) (227) (234) (234)

LOGTNAit1 0024 0023 0026 0025 0022 0022 0025 0023(234) (230) (198) (183) (219) (215) (186) (170)

LOGFAMSIZEit1 0016 0016 0016 0016 0016 0016 0016 0016(238) (238) (238) (238) (240) (239) (240) (239)

TURNOVERit1 0000 0000 0000 0000 0000 0000 0000 0000(128) (129) (128) (129) (126) (126) (126) (126)

AGEit1 0001 0001 0001 0001 0001 0001 0001 0001(064) (065) (064) (064) (060) (061) (059) (059)

EXPRATIOit1 0002 0002 0003 0003 0043 0043 0045 0045(003) (003) (004) (004) (054) (054) (055) (055)

TOTLOADit1 0002 0002 0002 0002 0002 0002 0002 0002(028) (029) (030) (029) (036) (036) (037) (037)

FLOWit1 0000 0000 0000 0000 0000 0000 0000 0000(164) (168) (168) (169) (164) (168) (168) (169)

FUNDRETit1 0030 0030 0029 0029 0031 0030 0030 0030(338) (338) (334) (334) (341) (341) (337) (337)

No of months 96 96 96 96 96 96 96 96





REFERENCES




































































(1)








i3 HMLt it t 1 T



t 1 T




FactorMeanreturn

SD ofreturn

Cross-correlations

VWRF SMB HML MOM12

VWRF 058 437 100 032 039 002SMB 017 290 100 016 030HML 034 263 100 015MOM12 096 388 100


Portfolio

CAPM

Alpha VWRF

1 (small) 004 0872 002 0913 001 0934 009 0925 (large) 007 091


Portfolio

3-Factor model

HML

4-Factor model


1 (small) 001 082 029 003 002 082 030 004 0032 003 084 027 001 009 085 030 000 0053 001 087 022 005 006 087 025 003 0064 006 087 018 005 013 087 020 004 0065 (large) 005 088 008 006 010 088 010 005 005








Xit 1 it i 1 N






Xit 1 it i 1 N


III Results











INTERCEPT 0056 0087 0080 0019 0026 0056 0049 0011(096) (163) (128) (030) (044) (105) (079) (018)

LOGTNAit1 0028 0028 0023 0020 0025 0025 0020 0018(302) (304) (242) (215) (275) (276) (216) (189)

LOGFAMSIZEit1 0007 0007 0007 0007 0007 0007 0007 0007(226) (227) (222) (221) (233) (234) (229) (228)

TURNOVERit1 0000 0000 0000 0000 0000 0000 0000 0000(085) (086) (085) (083) (077) (078) (077) (075)

AGEit1 0001 0001 0001 0001 0000 0000 0001 0001(062) (062) (064) (063) (052) (053) (055) (054)

EXPRATIOit1 0004 0004 0007 0007 0039 0038 0041 0041(011) (009) (018) (018) (097) (095) (104) (105)

TOTLOADit1 0003 0003 0003 0003 0003 0003 0003 0003(126) (125) (126) (129) (121) (120) (121) (125)

FLOWit1 0000 0000 0000 0000 0000 0000 0000 0000(050) (050) (051) (049) (047) (047) (048) (046)

LAGFUNDRETit1 0029 0028 0028 0029 0029 0029 0029 0029(600) (598) (600) (599) (603) (601) (603) (602)

No of months 444 444 444 444 444 444 444 444



















LOGTNAit1 0023 0026 0019 0021 0020 0022 0015 0017(259) (294) (222) (248) (216) (249) (176) (202)

LOGFAMSIZEit1 0006 0006 0006 0006 0006 0006 0006 0006(199) (203) (199) (205) (210) (215) (211) (217)

No of months 444 444 444 444 444 444 444 444




LOGTNAit1 0436 0440 0345 0312 0402 0406 0312 0280(326) (327) (240) (219) (307) (307) (220) (199)

LOGFAMSIZEit1 0088 0089 0088 0088 0090 0091 0090 0090(205) (206) (205) (207) (210) (212) (211) (212)

No of years 37 37 37 37 37 37 37 37




LOGTNAit1 0035 0034 0021 0019 0032 0032 0019 0016(241) (242) (151) (133) (222) (223) (132) (115)

LOGFAMSIZEit1 0002 0002 0002 0002 0002 0002 0002 0002(049) (050) (045) (044) (056) (058) (053) (052)

No of months 216 216 216 216 216 216 216 216




LOGTNAit1 0021 0022 0024 0022 0019 0019 0022 0019(184) (186) (193) (171) (164) (165) (174) (153)

LOGFAMSIZEit1 0012 0012 0012 0012 0012 0012 0012 0012(253) (253) (250) (250) (255) (255) (252) (252)

No of months 228 228 228 228 228 228 228 228





Portfolio





















INTERCEPT 0205 0257 0320 0263 0175 0226 0289 0231(127) (153) (244) (198) (108) (134) (220) (175)

INDnot SCG 0250 0252 0260 0262 0249 0251 0259 0261(154) (155) (162) (163) (153) (154) (161) (162)

LOGTNAit1 0058 0058 0063 0061 0055 0056 0060 0058(315) (317) (340) (329) (303) (305) (328) (317)


LOGFAMSIZEit1 0012 0012 0011 0011 0012 0012 0012 0012(257) (257) (254) (254) (259) (259) (257) (256)

TURNOVERit1 0000 0000 0000 0000 0000 0000 0000 0000(162) (161) (162) (161) (157) (156) (157) (157)

AGEit1 0001 0001 0001 0001 0001 0001 0001 0001(098) (097) (098) (097) (087) (087) (088) (087)

EXPRATIOit1 0018 0018 0019 0019 0062 0062 0064 0063(044) (043) (046) (044) (149) (148) (152) (151)

TOTLOADit1 0001 0001 0001 0001 0001 0001 0001 0001(054) (054) (052) (052) (048) (047) (046) (046)

FLOWit1 0000 0000 0000 0000 0000 0000 0000 0000(096) (095) (090) (092) (102) (102) (096) (099)

FUNDRETit1 0022 0022 0021 0021 0022 0022 0022 0022(345) (345) (343) (342) (348) (348) (345) (345)

No of months 228 228 228 228 228 228 228 228




INTERCEPT 0016 0068 0128 0068 0017 0035 0095 0034(019) (084) (172) (092) (020) (043) (127) (047)

INDLC 0119 0121 0119 0120 0112 0114 0112 0113(131) (133) (133) (134) (123) (125) (124) (126)

LOGTNAit1 0029 0029 0032 0029 0027 0027 0029 0027(222) (224) (245) (225) (202) (205) (226) (205)

LOGTNAit1 0031 0031 0031 0031 0030 0030 0030 0030INDLC (229) (232) (234) (236) (223) (226) (228) (230)

LOGFAMSIZEit1 0011 0011 0011 0011 0011 0011 0011 0011(254) (254) (251) (251) (258) (258) (255) (255)

TURNOVERit1 0000 0000 0000 0000 0000 0000 0000 0000(180) (179) (180) (179) (175) (175) (176) (175)

AGEit1 0001 0001 0001 0001 0001 0001 0001 0001(126) (125) (126) (125) (117) (117) (117) (116)

EXPRATIOit1 0020 0019 0021 0020 0065 0065 0066 0065(049) (048) (052) (050) (163) (162) (166) (164)

TOTLOADit1 0001 0001 0001 0001 0001 0001 0001 0001(035) (035) (034) (035) (030) (030) (029) (029)

FLOWit1 0000 0000 0000 0000 0000 0000 0000 0000(084) (082) (077) (079) (089) (088) (083) (085)

FUNDRETit1 0020 0020 0020 0020 0021 0020 0020 0020(310) (309) (307) (307) (313) (312) (310) (309)

No of months 228 228 228 228 228 228 228 228












LOGTNAit1 0051 0052 0056 0054 0048 0049 0053 0051(283) (286) (313) (302) (270) (272) (301) (289)


LOGFAMSIZEit1 0007 0007 0007 0007 0007 0007 0007 0007(070) (070) (069) (070) (070) (070) (069) (070)


No of months 228 228 228 228 228 228 228 228




LOGTNAit1 0067 0070 0071 0072 0064 0067 0068 0069(367) (383) (387) (391) (351) (367) (371) (376)


LOGFAMSIZEit1 0012 0012 0012 0012 0012 0012 0012 0012(113) (113) (111) (111) (115) (115) (113) (113)


No of months 228 228 228 228 228 228 228 228




































(1) (2) (3)

SMALLFUND 00759 00580 00584(949) (674) (679)

SMALLCAPFUND 00515 00340 00341(628) (386) (388)














(1) (2) (3)

SMALLFUND 00258 00143 00150(183) (094) (099)

SMALLCAPFUND 01190 01078 01079(821) (695) (696)














STOCKS


MULTI-MANAGER 00102 00190(175) (173)


















V Conclusion





INTERCEPT 0069 0018 0039 0008 0092 0042 0014 0032(066) (019) (025) (005) (088) (044) (009) (021)

MULTI-MANAGERi 0040 0040 0041 0041 0040 0041 0042 0042(222) (223) (230) (230) (226) (227) (234) (234)

LOGTNAit1 0024 0023 0026 0025 0022 0022 0025 0023(234) (230) (198) (183) (219) (215) (186) (170)

LOGFAMSIZEit1 0016 0016 0016 0016 0016 0016 0016 0016(238) (238) (238) (238) (240) (239) (240) (239)

TURNOVERit1 0000 0000 0000 0000 0000 0000 0000 0000(128) (129) (128) (129) (126) (126) (126) (126)

AGEit1 0001 0001 0001 0001 0001 0001 0001 0001(064) (065) (064) (064) (060) (061) (059) (059)

EXPRATIOit1 0002 0002 0003 0003 0043 0043 0045 0045(003) (003) (004) (004) (054) (054) (055) (055)

TOTLOADit1 0002 0002 0002 0002 0002 0002 0002 0002(028) (029) (030) (029) (036) (036) (037) (037)

FLOWit1 0000 0000 0000 0000 0000 0000 0000 0000(164) (168) (168) (169) (164) (168) (168) (169)

FUNDRETit1 0030 0030 0029 0029 0031 0030 0030 0030(338) (338) (334) (334) (341) (341) (337) (337)

No of months 96 96 96 96 96 96 96 96





REFERENCES































































i3 HMLt it t 1 T



t 1 T




FactorMeanreturn

SD ofreturn

Cross-correlations

VWRF SMB HML MOM12

VWRF 058 437 100 032 039 002SMB 017 290 100 016 030HML 034 263 100 015MOM12 096 388 100


Portfolio

CAPM

Alpha VWRF

1 (small) 004 0872 002 0913 001 0934 009 0925 (large) 007 091


Portfolio

3-Factor model

HML

4-Factor model


1 (small) 001 082 029 003 002 082 030 004 0032 003 084 027 001 009 085 030 000 0053 001 087 022 005 006 087 025 003 0064 006 087 018 005 013 087 020 004 0065 (large) 005 088 008 006 010 088 010 005 005








Xit 1 it i 1 N






Xit 1 it i 1 N


III Results











INTERCEPT 0056 0087 0080 0019 0026 0056 0049 0011(096) (163) (128) (030) (044) (105) (079) (018)

LOGTNAit1 0028 0028 0023 0020 0025 0025 0020 0018(302) (304) (242) (215) (275) (276) (216) (189)

LOGFAMSIZEit1 0007 0007 0007 0007 0007 0007 0007 0007(226) (227) (222) (221) (233) (234) (229) (228)

TURNOVERit1 0000 0000 0000 0000 0000 0000 0000 0000(085) (086) (085) (083) (077) (078) (077) (075)

AGEit1 0001 0001 0001 0001 0000 0000 0001 0001(062) (062) (064) (063) (052) (053) (055) (054)

EXPRATIOit1 0004 0004 0007 0007 0039 0038 0041 0041(011) (009) (018) (018) (097) (095) (104) (105)

TOTLOADit1 0003 0003 0003 0003 0003 0003 0003 0003(126) (125) (126) (129) (121) (120) (121) (125)

FLOWit1 0000 0000 0000 0000 0000 0000 0000 0000(050) (050) (051) (049) (047) (047) (048) (046)

LAGFUNDRETit1 0029 0028 0028 0029 0029 0029 0029 0029(600) (598) (600) (599) (603) (601) (603) (602)

No of months 444 444 444 444 444 444 444 444



















LOGTNAit1 0023 0026 0019 0021 0020 0022 0015 0017(259) (294) (222) (248) (216) (249) (176) (202)

LOGFAMSIZEit1 0006 0006 0006 0006 0006 0006 0006 0006(199) (203) (199) (205) (210) (215) (211) (217)

No of months 444 444 444 444 444 444 444 444




LOGTNAit1 0436 0440 0345 0312 0402 0406 0312 0280(326) (327) (240) (219) (307) (307) (220) (199)

LOGFAMSIZEit1 0088 0089 0088 0088 0090 0091 0090 0090(205) (206) (205) (207) (210) (212) (211) (212)

No of years 37 37 37 37 37 37 37 37




LOGTNAit1 0035 0034 0021 0019 0032 0032 0019 0016(241) (242) (151) (133) (222) (223) (132) (115)

LOGFAMSIZEit1 0002 0002 0002 0002 0002 0002 0002 0002(049) (050) (045) (044) (056) (058) (053) (052)

No of months 216 216 216 216 216 216 216 216




LOGTNAit1 0021 0022 0024 0022 0019 0019 0022 0019(184) (186) (193) (171) (164) (165) (174) (153)

LOGFAMSIZEit1 0012 0012 0012 0012 0012 0012 0012 0012(253) (253) (250) (250) (255) (255) (252) (252)

No of months 228 228 228 228 228 228 228 228





Portfolio





















INTERCEPT 0205 0257 0320 0263 0175 0226 0289 0231(127) (153) (244) (198) (108) (134) (220) (175)

INDnot SCG 0250 0252 0260 0262 0249 0251 0259 0261(154) (155) (162) (163) (153) (154) (161) (162)

LOGTNAit1 0058 0058 0063 0061 0055 0056 0060 0058(315) (317) (340) (329) (303) (305) (328) (317)


LOGFAMSIZEit1 0012 0012 0011 0011 0012 0012 0012 0012(257) (257) (254) (254) (259) (259) (257) (256)

TURNOVERit1 0000 0000 0000 0000 0000 0000 0000 0000(162) (161) (162) (161) (157) (156) (157) (157)

AGEit1 0001 0001 0001 0001 0001 0001 0001 0001(098) (097) (098) (097) (087) (087) (088) (087)

EXPRATIOit1 0018 0018 0019 0019 0062 0062 0064 0063(044) (043) (046) (044) (149) (148) (152) (151)

TOTLOADit1 0001 0001 0001 0001 0001 0001 0001 0001(054) (054) (052) (052) (048) (047) (046) (046)

FLOWit1 0000 0000 0000 0000 0000 0000 0000 0000(096) (095) (090) (092) (102) (102) (096) (099)

FUNDRETit1 0022 0022 0021 0021 0022 0022 0022 0022(345) (345) (343) (342) (348) (348) (345) (345)

No of months 228 228 228 228 228 228 228 228




INTERCEPT 0016 0068 0128 0068 0017 0035 0095 0034(019) (084) (172) (092) (020) (043) (127) (047)

INDLC 0119 0121 0119 0120 0112 0114 0112 0113(131) (133) (133) (134) (123) (125) (124) (126)

LOGTNAit1 0029 0029 0032 0029 0027 0027 0029 0027(222) (224) (245) (225) (202) (205) (226) (205)

LOGTNAit1 0031 0031 0031 0031 0030 0030 0030 0030INDLC (229) (232) (234) (236) (223) (226) (228) (230)

LOGFAMSIZEit1 0011 0011 0011 0011 0011 0011 0011 0011(254) (254) (251) (251) (258) (258) (255) (255)

TURNOVERit1 0000 0000 0000 0000 0000 0000 0000 0000(180) (179) (180) (179) (175) (175) (176) (175)

AGEit1 0001 0001 0001 0001 0001 0001 0001 0001(126) (125) (126) (125) (117) (117) (117) (116)

EXPRATIOit1 0020 0019 0021 0020 0065 0065 0066 0065(049) (048) (052) (050) (163) (162) (166) (164)

TOTLOADit1 0001 0001 0001 0001 0001 0001 0001 0001(035) (035) (034) (035) (030) (030) (029) (029)

FLOWit1 0000 0000 0000 0000 0000 0000 0000 0000(084) (082) (077) (079) (089) (088) (083) (085)

FUNDRETit1 0020 0020 0020 0020 0021 0020 0020 0020(310) (309) (307) (307) (313) (312) (310) (309)

No of months 228 228 228 228 228 228 228 228












LOGTNAit1 0051 0052 0056 0054 0048 0049 0053 0051(283) (286) (313) (302) (270) (272) (301) (289)


LOGFAMSIZEit1 0007 0007 0007 0007 0007 0007 0007 0007(070) (070) (069) (070) (070) (070) (069) (070)


No of months 228 228 228 228 228 228 228 228




LOGTNAit1 0067 0070 0071 0072 0064 0067 0068 0069(367) (383) (387) (391) (351) (367) (371) (376)


LOGFAMSIZEit1 0012 0012 0012 0012 0012 0012 0012 0012(113) (113) (111) (111) (115) (115) (113) (113)


No of months 228 228 228 228 228 228 228 228




































(1) (2) (3)

SMALLFUND 00759 00580 00584(949) (674) (679)

SMALLCAPFUND 00515 00340 00341(628) (386) (388)














(1) (2) (3)

SMALLFUND 00258 00143 00150(183) (094) (099)

SMALLCAPFUND 01190 01078 01079(821) (695) (696)














STOCKS


MULTI-MANAGER 00102 00190(175) (173)


















V Conclusion





INTERCEPT 0069 0018 0039 0008 0092 0042 0014 0032(066) (019) (025) (005) (088) (044) (009) (021)

MULTI-MANAGERi 0040 0040 0041 0041 0040 0041 0042 0042(222) (223) (230) (230) (226) (227) (234) (234)

LOGTNAit1 0024 0023 0026 0025 0022 0022 0025 0023(234) (230) (198) (183) (219) (215) (186) (170)

LOGFAMSIZEit1 0016 0016 0016 0016 0016 0016 0016 0016(238) (238) (238) (238) (240) (239) (240) (239)

TURNOVERit1 0000 0000 0000 0000 0000 0000 0000 0000(128) (129) (128) (129) (126) (126) (126) (126)

AGEit1 0001 0001 0001 0001 0001 0001 0001 0001(064) (065) (064) (064) (060) (061) (059) (059)

EXPRATIOit1 0002 0002 0003 0003 0043 0043 0045 0045(003) (003) (004) (004) (054) (054) (055) (055)

TOTLOADit1 0002 0002 0002 0002 0002 0002 0002 0002(028) (029) (030) (029) (036) (036) (037) (037)

FLOWit1 0000 0000 0000 0000 0000 0000 0000 0000(164) (168) (168) (169) (164) (168) (168) (169)

FUNDRETit1 0030 0030 0029 0029 0031 0030 0030 0030(338) (338) (334) (334) (341) (341) (337) (337)

No of months 96 96 96 96 96 96 96 96





REFERENCES

































































Xit 1 it i 1 N






Xit 1 it i 1 N


III Results











INTERCEPT 0056 0087 0080 0019 0026 0056 0049 0011(096) (163) (128) (030) (044) (105) (079) (018)

LOGTNAit1 0028 0028 0023 0020 0025 0025 0020 0018(302) (304) (242) (215) (275) (276) (216) (189)

LOGFAMSIZEit1 0007 0007 0007 0007 0007 0007 0007 0007(226) (227) (222) (221) (233) (234) (229) (228)

TURNOVERit1 0000 0000 0000 0000 0000 0000 0000 0000(085) (086) (085) (083) (077) (078) (077) (075)

AGEit1 0001 0001 0001 0001 0000 0000 0001 0001(062) (062) (064) (063) (052) (053) (055) (054)

EXPRATIOit1 0004 0004 0007 0007 0039 0038 0041 0041(011) (009) (018) (018) (097) (095) (104) (105)

TOTLOADit1 0003 0003 0003 0003 0003 0003 0003 0003(126) (125) (126) (129) (121) (120) (121) (125)

FLOWit1 0000 0000 0000 0000 0000 0000 0000 0000(050) (050) (051) (049) (047) (047) (048) (046)

LAGFUNDRETit1 0029 0028 0028 0029 0029 0029 0029 0029(600) (598) (600) (599) (603) (601) (603) (602)

No of months 444 444 444 444 444 444 444 444



















LOGTNAit1 0023 0026 0019 0021 0020 0022 0015 0017(259) (294) (222) (248) (216) (249) (176) (202)

LOGFAMSIZEit1 0006 0006 0006 0006 0006 0006 0006 0006(199) (203) (199) (205) (210) (215) (211) (217)

No of months 444 444 444 444 444 444 444 444




LOGTNAit1 0436 0440 0345 0312 0402 0406 0312 0280(326) (327) (240) (219) (307) (307) (220) (199)

LOGFAMSIZEit1 0088 0089 0088 0088 0090 0091 0090 0090(205) (206) (205) (207) (210) (212) (211) (212)

No of years 37 37 37 37 37 37 37 37




LOGTNAit1 0035 0034 0021 0019 0032 0032 0019 0016(241) (242) (151) (133) (222) (223) (132) (115)

LOGFAMSIZEit1 0002 0002 0002 0002 0002 0002 0002 0002(049) (050) (045) (044) (056) (058) (053) (052)

No of months 216 216 216 216 216 216 216 216




LOGTNAit1 0021 0022 0024 0022 0019 0019 0022 0019(184) (186) (193) (171) (164) (165) (174) (153)

LOGFAMSIZEit1 0012 0012 0012 0012 0012 0012 0012 0012(253) (253) (250) (250) (255) (255) (252) (252)

No of months 228 228 228 228 228 228 228 228





Portfolio





















INTERCEPT 0205 0257 0320 0263 0175 0226 0289 0231(127) (153) (244) (198) (108) (134) (220) (175)

INDnot SCG 0250 0252 0260 0262 0249 0251 0259 0261(154) (155) (162) (163) (153) (154) (161) (162)

LOGTNAit1 0058 0058 0063 0061 0055 0056 0060 0058(315) (317) (340) (329) (303) (305) (328) (317)


LOGFAMSIZEit1 0012 0012 0011 0011 0012 0012 0012 0012(257) (257) (254) (254) (259) (259) (257) (256)

TURNOVERit1 0000 0000 0000 0000 0000 0000 0000 0000(162) (161) (162) (161) (157) (156) (157) (157)

AGEit1 0001 0001 0001 0001 0001 0001 0001 0001(098) (097) (098) (097) (087) (087) (088) (087)

EXPRATIOit1 0018 0018 0019 0019 0062 0062 0064 0063(044) (043) (046) (044) (149) (148) (152) (151)

TOTLOADit1 0001 0001 0001 0001 0001 0001 0001 0001(054) (054) (052) (052) (048) (047) (046) (046)

FLOWit1 0000 0000 0000 0000 0000 0000 0000 0000(096) (095) (090) (092) (102) (102) (096) (099)

FUNDRETit1 0022 0022 0021 0021 0022 0022 0022 0022(345) (345) (343) (342) (348) (348) (345) (345)

No of months 228 228 228 228 228 228 228 228




INTERCEPT 0016 0068 0128 0068 0017 0035 0095 0034(019) (084) (172) (092) (020) (043) (127) (047)

INDLC 0119 0121 0119 0120 0112 0114 0112 0113(131) (133) (133) (134) (123) (125) (124) (126)

LOGTNAit1 0029 0029 0032 0029 0027 0027 0029 0027(222) (224) (245) (225) (202) (205) (226) (205)

LOGTNAit1 0031 0031 0031 0031 0030 0030 0030 0030INDLC (229) (232) (234) (236) (223) (226) (228) (230)

LOGFAMSIZEit1 0011 0011 0011 0011 0011 0011 0011 0011(254) (254) (251) (251) (258) (258) (255) (255)

TURNOVERit1 0000 0000 0000 0000 0000 0000 0000 0000(180) (179) (180) (179) (175) (175) (176) (175)

AGEit1 0001 0001 0001 0001 0001 0001 0001 0001(126) (125) (126) (125) (117) (117) (117) (116)

EXPRATIOit1 0020 0019 0021 0020 0065 0065 0066 0065(049) (048) (052) (050) (163) (162) (166) (164)

TOTLOADit1 0001 0001 0001 0001 0001 0001 0001 0001(035) (035) (034) (035) (030) (030) (029) (029)

FLOWit1 0000 0000 0000 0000 0000 0000 0000 0000(084) (082) (077) (079) (089) (088) (083) (085)

FUNDRETit1 0020 0020 0020 0020 0021 0020 0020 0020(310) (309) (307) (307) (313) (312) (310) (309)

No of months 228 228 228 228 228 228 228 228












LOGTNAit1 0051 0052 0056 0054 0048 0049 0053 0051(283) (286) (313) (302) (270) (272) (301) (289)


LOGFAMSIZEit1 0007 0007 0007 0007 0007 0007 0007 0007(070) (070) (069) (070) (070) (070) (069) (070)


No of months 228 228 228 228 228 228 228 228




LOGTNAit1 0067 0070 0071 0072 0064 0067 0068 0069(367) (383) (387) (391) (351) (367) (371) (376)


LOGFAMSIZEit1 0012 0012 0012 0012 0012 0012 0012 0012(113) (113) (111) (111) (115) (115) (113) (113)


No of months 228 228 228 228 228 228 228 228




































(1) (2) (3)

SMALLFUND 00759 00580 00584(949) (674) (679)

SMALLCAPFUND 00515 00340 00341(628) (386) (388)














(1) (2) (3)

SMALLFUND 00258 00143 00150(183) (094) (099)

SMALLCAPFUND 01190 01078 01079(821) (695) (696)














STOCKS


MULTI-MANAGER 00102 00190(175) (173)


















V Conclusion





INTERCEPT 0069 0018 0039 0008 0092 0042 0014 0032(066) (019) (025) (005) (088) (044) (009) (021)

MULTI-MANAGERi 0040 0040 0041 0041 0040 0041 0042 0042(222) (223) (230) (230) (226) (227) (234) (234)

LOGTNAit1 0024 0023 0026 0025 0022 0022 0025 0023(234) (230) (198) (183) (219) (215) (186) (170)

LOGFAMSIZEit1 0016 0016 0016 0016 0016 0016 0016 0016(238) (238) (238) (238) (240) (239) (240) (239)

TURNOVERit1 0000 0000 0000 0000 0000 0000 0000 0000(128) (129) (128) (129) (126) (126) (126) (126)

AGEit1 0001 0001 0001 0001 0001 0001 0001 0001(064) (065) (064) (064) (060) (061) (059) (059)

EXPRATIOit1 0002 0002 0003 0003 0043 0043 0045 0045(003) (003) (004) (004) (054) (054) (055) (055)

TOTLOADit1 0002 0002 0002 0002 0002 0002 0002 0002(028) (029) (030) (029) (036) (036) (037) (037)

FLOWit1 0000 0000 0000 0000 0000 0000 0000 0000(164) (168) (168) (169) (164) (168) (168) (169)

FUNDRETit1 0030 0030 0029 0029 0031 0030 0030 0030(338) (338) (334) (334) (341) (341) (337) (337)

No of months 96 96 96 96 96 96 96 96





REFERENCES



































































INTERCEPT 0056 0087 0080 0019 0026 0056 0049 0011(096) (163) (128) (030) (044) (105) (079) (018)

LOGTNAit1 0028 0028 0023 0020 0025 0025 0020 0018(302) (304) (242) (215) (275) (276) (216) (189)

LOGFAMSIZEit1 0007 0007 0007 0007 0007 0007 0007 0007(226) (227) (222) (221) (233) (234) (229) (228)

TURNOVERit1 0000 0000 0000 0000 0000 0000 0000 0000(085) (086) (085) (083) (077) (078) (077) (075)

AGEit1 0001 0001 0001 0001 0000 0000 0001 0001(062) (062) (064) (063) (052) (053) (055) (054)

EXPRATIOit1 0004 0004 0007 0007 0039 0038 0041 0041(011) (009) (018) (018) (097) (095) (104) (105)

TOTLOADit1 0003 0003 0003 0003 0003 0003 0003 0003(126) (125) (126) (129) (121) (120) (121) (125)

FLOWit1 0000 0000 0000 0000 0000 0000 0000 0000(050) (050) (051) (049) (047) (047) (048) (046)

LAGFUNDRETit1 0029 0028 0028 0029 0029 0029 0029 0029(600) (598) (600) (599) (603) (601) (603) (602)

No of months 444 444 444 444 444 444 444 444



















LOGTNAit1 0023 0026 0019 0021 0020 0022 0015 0017(259) (294) (222) (248) (216) (249) (176) (202)

LOGFAMSIZEit1 0006 0006 0006 0006 0006 0006 0006 0006(199) (203) (199) (205) (210) (215) (211) (217)

No of months 444 444 444 444 444 444 444 444




LOGTNAit1 0436 0440 0345 0312 0402 0406 0312 0280(326) (327) (240) (219) (307) (307) (220) (199)

LOGFAMSIZEit1 0088 0089 0088 0088 0090 0091 0090 0090(205) (206) (205) (207) (210) (212) (211) (212)

No of years 37 37 37 37 37 37 37 37




LOGTNAit1 0035 0034 0021 0019 0032 0032 0019 0016(241) (242) (151) (133) (222) (223) (132) (115)

LOGFAMSIZEit1 0002 0002 0002 0002 0002 0002 0002 0002(049) (050) (045) (044) (056) (058) (053) (052)

No of months 216 216 216 216 216 216 216 216




LOGTNAit1 0021 0022 0024 0022 0019 0019 0022 0019(184) (186) (193) (171) (164) (165) (174) (153)

LOGFAMSIZEit1 0012 0012 0012 0012 0012 0012 0012 0012(253) (253) (250) (250) (255) (255) (252) (252)

No of months 228 228 228 228 228 228 228 228





Portfolio





















INTERCEPT 0205 0257 0320 0263 0175 0226 0289 0231(127) (153) (244) (198) (108) (134) (220) (175)

INDnot SCG 0250 0252 0260 0262 0249 0251 0259 0261(154) (155) (162) (163) (153) (154) (161) (162)

LOGTNAit1 0058 0058 0063 0061 0055 0056 0060 0058(315) (317) (340) (329) (303) (305) (328) (317)


LOGFAMSIZEit1 0012 0012 0011 0011 0012 0012 0012 0012(257) (257) (254) (254) (259) (259) (257) (256)

TURNOVERit1 0000 0000 0000 0000 0000 0000 0000 0000(162) (161) (162) (161) (157) (156) (157) (157)

AGEit1 0001 0001 0001 0001 0001 0001 0001 0001(098) (097) (098) (097) (087) (087) (088) (087)

EXPRATIOit1 0018 0018 0019 0019 0062 0062 0064 0063(044) (043) (046) (044) (149) (148) (152) (151)

TOTLOADit1 0001 0001 0001 0001 0001 0001 0001 0001(054) (054) (052) (052) (048) (047) (046) (046)

FLOWit1 0000 0000 0000 0000 0000 0000 0000 0000(096) (095) (090) (092) (102) (102) (096) (099)

FUNDRETit1 0022 0022 0021 0021 0022 0022 0022 0022(345) (345) (343) (342) (348) (348) (345) (345)

No of months 228 228 228 228 228 228 228 228




INTERCEPT 0016 0068 0128 0068 0017 0035 0095 0034(019) (084) (172) (092) (020) (043) (127) (047)

INDLC 0119 0121 0119 0120 0112 0114 0112 0113(131) (133) (133) (134) (123) (125) (124) (126)

LOGTNAit1 0029 0029 0032 0029 0027 0027 0029 0027(222) (224) (245) (225) (202) (205) (226) (205)

LOGTNAit1 0031 0031 0031 0031 0030 0030 0030 0030INDLC (229) (232) (234) (236) (223) (226) (228) (230)

LOGFAMSIZEit1 0011 0011 0011 0011 0011 0011 0011 0011(254) (254) (251) (251) (258) (258) (255) (255)

TURNOVERit1 0000 0000 0000 0000 0000 0000 0000 0000(180) (179) (180) (179) (175) (175) (176) (175)

AGEit1 0001 0001 0001 0001 0001 0001 0001 0001(126) (125) (126) (125) (117) (117) (117) (116)

EXPRATIOit1 0020 0019 0021 0020 0065 0065 0066 0065(049) (048) (052) (050) (163) (162) (166) (164)

TOTLOADit1 0001 0001 0001 0001 0001 0001 0001 0001(035) (035) (034) (035) (030) (030) (029) (029)

FLOWit1 0000 0000 0000 0000 0000 0000 0000 0000(084) (082) (077) (079) (089) (088) (083) (085)

FUNDRETit1 0020 0020 0020 0020 0021 0020 0020 0020(310) (309) (307) (307) (313) (312) (310) (309)

No of months 228 228 228 228 228 228 228 228












LOGTNAit1 0051 0052 0056 0054 0048 0049 0053 0051(283) (286) (313) (302) (270) (272) (301) (289)


LOGFAMSIZEit1 0007 0007 0007 0007 0007 0007 0007 0007(070) (070) (069) (070) (070) (070) (069) (070)


No of months 228 228 228 228 228 228 228 228




LOGTNAit1 0067 0070 0071 0072 0064 0067 0068 0069(367) (383) (387) (391) (351) (367) (371) (376)


LOGFAMSIZEit1 0012 0012 0012 0012 0012 0012 0012 0012(113) (113) (111) (111) (115) (115) (113) (113)


No of months 228 228 228 228 228 228 228 228




































(1) (2) (3)

SMALLFUND 00759 00580 00584(949) (674) (679)

SMALLCAPFUND 00515 00340 00341(628) (386) (388)














(1) (2) (3)

SMALLFUND 00258 00143 00150(183) (094) (099)

SMALLCAPFUND 01190 01078 01079(821) (695) (696)














STOCKS


MULTI-MANAGER 00102 00190(175) (173)


















V Conclusion





INTERCEPT 0069 0018 0039 0008 0092 0042 0014 0032(066) (019) (025) (005) (088) (044) (009) (021)

MULTI-MANAGERi 0040 0040 0041 0041 0040 0041 0042 0042(222) (223) (230) (230) (226) (227) (234) (234)

LOGTNAit1 0024 0023 0026 0025 0022 0022 0025 0023(234) (230) (198) (183) (219) (215) (186) (170)

LOGFAMSIZEit1 0016 0016 0016 0016 0016 0016 0016 0016(238) (238) (238) (238) (240) (239) (240) (239)

TURNOVERit1 0000 0000 0000 0000 0000 0000 0000 0000(128) (129) (128) (129) (126) (126) (126) (126)

AGEit1 0001 0001 0001 0001 0001 0001 0001 0001(064) (065) (064) (064) (060) (061) (059) (059)

EXPRATIOit1 0002 0002 0003 0003 0043 0043 0045 0045(003) (003) (004) (004) (054) (054) (055) (055)

TOTLOADit1 0002 0002 0002 0002 0002 0002 0002 0002(028) (029) (030) (029) (036) (036) (037) (037)

FLOWit1 0000 0000 0000 0000 0000 0000 0000 0000(164) (168) (168) (169) (164) (168) (168) (169)

FUNDRETit1 0030 0030 0029 0029 0031 0030 0030 0030(338) (338) (334) (334) (341) (341) (337) (337)

No of months 96 96 96 96 96 96 96 96





REFERENCES












































































LOGTNAit1 0023 0026 0019 0021 0020 0022 0015 0017(259) (294) (222) (248) (216) (249) (176) (202)

LOGFAMSIZEit1 0006 0006 0006 0006 0006 0006 0006 0006(199) (203) (199) (205) (210) (215) (211) (217)

No of months 444 444 444 444 444 444 444 444




LOGTNAit1 0436 0440 0345 0312 0402 0406 0312 0280(326) (327) (240) (219) (307) (307) (220) (199)

LOGFAMSIZEit1 0088 0089 0088 0088 0090 0091 0090 0090(205) (206) (205) (207) (210) (212) (211) (212)

No of years 37 37 37 37 37 37 37 37




LOGTNAit1 0035 0034 0021 0019 0032 0032 0019 0016(241) (242) (151) (133) (222) (223) (132) (115)

LOGFAMSIZEit1 0002 0002 0002 0002 0002 0002 0002 0002(049) (050) (045) (044) (056) (058) (053) (052)

No of months 216 216 216 216 216 216 216 216




LOGTNAit1 0021 0022 0024 0022 0019 0019 0022 0019(184) (186) (193) (171) (164) (165) (174) (153)

LOGFAMSIZEit1 0012 0012 0012 0012 0012 0012 0012 0012(253) (253) (250) (250) (255) (255) (252) (252)

No of months 228 228 228 228 228 228 228 228





Portfolio





















INTERCEPT 0205 0257 0320 0263 0175 0226 0289 0231(127) (153) (244) (198) (108) (134) (220) (175)

INDnot SCG 0250 0252 0260 0262 0249 0251 0259 0261(154) (155) (162) (163) (153) (154) (161) (162)

LOGTNAit1 0058 0058 0063 0061 0055 0056 0060 0058(315) (317) (340) (329) (303) (305) (328) (317)


LOGFAMSIZEit1 0012 0012 0011 0011 0012 0012 0012 0012(257) (257) (254) (254) (259) (259) (257) (256)

TURNOVERit1 0000 0000 0000 0000 0000 0000 0000 0000(162) (161) (162) (161) (157) (156) (157) (157)

AGEit1 0001 0001 0001 0001 0001 0001 0001 0001(098) (097) (098) (097) (087) (087) (088) (087)

EXPRATIOit1 0018 0018 0019 0019 0062 0062 0064 0063(044) (043) (046) (044) (149) (148) (152) (151)

TOTLOADit1 0001 0001 0001 0001 0001 0001 0001 0001(054) (054) (052) (052) (048) (047) (046) (046)

FLOWit1 0000 0000 0000 0000 0000 0000 0000 0000(096) (095) (090) (092) (102) (102) (096) (099)

FUNDRETit1 0022 0022 0021 0021 0022 0022 0022 0022(345) (345) (343) (342) (348) (348) (345) (345)

No of months 228 228 228 228 228 228 228 228




INTERCEPT 0016 0068 0128 0068 0017 0035 0095 0034(019) (084) (172) (092) (020) (043) (127) (047)

INDLC 0119 0121 0119 0120 0112 0114 0112 0113(131) (133) (133) (134) (123) (125) (124) (126)

LOGTNAit1 0029 0029 0032 0029 0027 0027 0029 0027(222) (224) (245) (225) (202) (205) (226) (205)

LOGTNAit1 0031 0031 0031 0031 0030 0030 0030 0030INDLC (229) (232) (234) (236) (223) (226) (228) (230)

LOGFAMSIZEit1 0011 0011 0011 0011 0011 0011 0011 0011(254) (254) (251) (251) (258) (258) (255) (255)

TURNOVERit1 0000 0000 0000 0000 0000 0000 0000 0000(180) (179) (180) (179) (175) (175) (176) (175)

AGEit1 0001 0001 0001 0001 0001 0001 0001 0001(126) (125) (126) (125) (117) (117) (117) (116)

EXPRATIOit1 0020 0019 0021 0020 0065 0065 0066 0065(049) (048) (052) (050) (163) (162) (166) (164)

TOTLOADit1 0001 0001 0001 0001 0001 0001 0001 0001(035) (035) (034) (035) (030) (030) (029) (029)

FLOWit1 0000 0000 0000 0000 0000 0000 0000 0000(084) (082) (077) (079) (089) (088) (083) (085)

FUNDRETit1 0020 0020 0020 0020 0021 0020 0020 0020(310) (309) (307) (307) (313) (312) (310) (309)

No of months 228 228 228 228 228 228 228 228












LOGTNAit1 0051 0052 0056 0054 0048 0049 0053 0051(283) (286) (313) (302) (270) (272) (301) (289)


LOGFAMSIZEit1 0007 0007 0007 0007 0007 0007 0007 0007(070) (070) (069) (070) (070) (070) (069) (070)


No of months 228 228 228 228 228 228 228 228




LOGTNAit1 0067 0070 0071 0072 0064 0067 0068 0069(367) (383) (387) (391) (351) (367) (371) (376)


LOGFAMSIZEit1 0012 0012 0012 0012 0012 0012 0012 0012(113) (113) (111) (111) (115) (115) (113) (113)


No of months 228 228 228 228 228 228 228 228




































(1) (2) (3)

SMALLFUND 00759 00580 00584(949) (674) (679)

SMALLCAPFUND 00515 00340 00341(628) (386) (388)














(1) (2) (3)

SMALLFUND 00258 00143 00150(183) (094) (099)

SMALLCAPFUND 01190 01078 01079(821) (695) (696)














STOCKS


MULTI-MANAGER 00102 00190(175) (173)


















V Conclusion





INTERCEPT 0069 0018 0039 0008 0092 0042 0014 0032(066) (019) (025) (005) (088) (044) (009) (021)

MULTI-MANAGERi 0040 0040 0041 0041 0040 0041 0042 0042(222) (223) (230) (230) (226) (227) (234) (234)

LOGTNAit1 0024 0023 0026 0025 0022 0022 0025 0023(234) (230) (198) (183) (219) (215) (186) (170)

LOGFAMSIZEit1 0016 0016 0016 0016 0016 0016 0016 0016(238) (238) (238) (238) (240) (239) (240) (239)

TURNOVERit1 0000 0000 0000 0000 0000 0000 0000 0000(128) (129) (128) (129) (126) (126) (126) (126)

AGEit1 0001 0001 0001 0001 0001 0001 0001 0001(064) (065) (064) (064) (060) (061) (059) (059)

EXPRATIOit1 0002 0002 0003 0003 0043 0043 0045 0045(003) (003) (004) (004) (054) (054) (055) (055)

TOTLOADit1 0002 0002 0002 0002 0002 0002 0002 0002(028) (029) (030) (029) (036) (036) (037) (037)

FLOWit1 0000 0000 0000 0000 0000 0000 0000 0000(164) (168) (168) (169) (164) (168) (168) (169)

FUNDRETit1 0030 0030 0029 0029 0031 0030 0030 0030(338) (338) (334) (334) (341) (341) (337) (337)

No of months 96 96 96 96 96 96 96 96





REFERENCES






























































Portfolio





















INTERCEPT 0205 0257 0320 0263 0175 0226 0289 0231(127) (153) (244) (198) (108) (134) (220) (175)

INDnot SCG 0250 0252 0260 0262 0249 0251 0259 0261(154) (155) (162) (163) (153) (154) (161) (162)

LOGTNAit1 0058 0058 0063 0061 0055 0056 0060 0058(315) (317) (340) (329) (303) (305) (328) (317)


LOGFAMSIZEit1 0012 0012 0011 0011 0012 0012 0012 0012(257) (257) (254) (254) (259) (259) (257) (256)

TURNOVERit1 0000 0000 0000 0000 0000 0000 0000 0000(162) (161) (162) (161) (157) (156) (157) (157)

AGEit1 0001 0001 0001 0001 0001 0001 0001 0001(098) (097) (098) (097) (087) (087) (088) (087)

EXPRATIOit1 0018 0018 0019 0019 0062 0062 0064 0063(044) (043) (046) (044) (149) (148) (152) (151)

TOTLOADit1 0001 0001 0001 0001 0001 0001 0001 0001(054) (054) (052) (052) (048) (047) (046) (046)

FLOWit1 0000 0000 0000 0000 0000 0000 0000 0000(096) (095) (090) (092) (102) (102) (096) (099)

FUNDRETit1 0022 0022 0021 0021 0022 0022 0022 0022(345) (345) (343) (342) (348) (348) (345) (345)

No of months 228 228 228 228 228 228 228 228




INTERCEPT 0016 0068 0128 0068 0017 0035 0095 0034(019) (084) (172) (092) (020) (043) (127) (047)

INDLC 0119 0121 0119 0120 0112 0114 0112 0113(131) (133) (133) (134) (123) (125) (124) (126)

LOGTNAit1 0029 0029 0032 0029 0027 0027 0029 0027(222) (224) (245) (225) (202) (205) (226) (205)

LOGTNAit1 0031 0031 0031 0031 0030 0030 0030 0030INDLC (229) (232) (234) (236) (223) (226) (228) (230)

LOGFAMSIZEit1 0011 0011 0011 0011 0011 0011 0011 0011(254) (254) (251) (251) (258) (258) (255) (255)

TURNOVERit1 0000 0000 0000 0000 0000 0000 0000 0000(180) (179) (180) (179) (175) (175) (176) (175)

AGEit1 0001 0001 0001 0001 0001 0001 0001 0001(126) (125) (126) (125) (117) (117) (117) (116)

EXPRATIOit1 0020 0019 0021 0020 0065 0065 0066 0065(049) (048) (052) (050) (163) (162) (166) (164)

TOTLOADit1 0001 0001 0001 0001 0001 0001 0001 0001(035) (035) (034) (035) (030) (030) (029) (029)

FLOWit1 0000 0000 0000 0000 0000 0000 0000 0000(084) (082) (077) (079) (089) (088) (083) (085)

FUNDRETit1 0020 0020 0020 0020 0021 0020 0020 0020(310) (309) (307) (307) (313) (312) (310) (309)

No of months 228 228 228 228 228 228 228 228












LOGTNAit1 0051 0052 0056 0054 0048 0049 0053 0051(283) (286) (313) (302) (270) (272) (301) (289)


LOGFAMSIZEit1 0007 0007 0007 0007 0007 0007 0007 0007(070) (070) (069) (070) (070) (070) (069) (070)


No of months 228 228 228 228 228 228 228 228




LOGTNAit1 0067 0070 0071 0072 0064 0067 0068 0069(367) (383) (387) (391) (351) (367) (371) (376)


LOGFAMSIZEit1 0012 0012 0012 0012 0012 0012 0012 0012(113) (113) (111) (111) (115) (115) (113) (113)


No of months 228 228 228 228 228 228 228 228




































(1) (2) (3)

SMALLFUND 00759 00580 00584(949) (674) (679)

SMALLCAPFUND 00515 00340 00341(628) (386) (388)














(1) (2) (3)

SMALLFUND 00258 00143 00150(183) (094) (099)

SMALLCAPFUND 01190 01078 01079(821) (695) (696)














STOCKS


MULTI-MANAGER 00102 00190(175) (173)


















V Conclusion





INTERCEPT 0069 0018 0039 0008 0092 0042 0014 0032(066) (019) (025) (005) (088) (044) (009) (021)

MULTI-MANAGERi 0040 0040 0041 0041 0040 0041 0042 0042(222) (223) (230) (230) (226) (227) (234) (234)

LOGTNAit1 0024 0023 0026 0025 0022 0022 0025 0023(234) (230) (198) (183) (219) (215) (186) (170)

LOGFAMSIZEit1 0016 0016 0016 0016 0016 0016 0016 0016(238) (238) (238) (238) (240) (239) (240) (239)

TURNOVERit1 0000 0000 0000 0000 0000 0000 0000 0000(128) (129) (128) (129) (126) (126) (126) (126)

AGEit1 0001 0001 0001 0001 0001 0001 0001 0001(064) (065) (064) (064) (060) (061) (059) (059)

EXPRATIOit1 0002 0002 0003 0003 0043 0043 0045 0045(003) (003) (004) (004) (054) (054) (055) (055)

TOTLOADit1 0002 0002 0002 0002 0002 0002 0002 0002(028) (029) (030) (029) (036) (036) (037) (037)

FLOWit1 0000 0000 0000 0000 0000 0000 0000 0000(164) (168) (168) (169) (164) (168) (168) (169)

FUNDRETit1 0030 0030 0029 0029 0031 0030 0030 0030(338) (338) (334) (334) (341) (341) (337) (337)

No of months 96 96 96 96 96 96 96 96





REFERENCES










































































INTERCEPT 0205 0257 0320 0263 0175 0226 0289 0231(127) (153) (244) (198) (108) (134) (220) (175)

INDnot SCG 0250 0252 0260 0262 0249 0251 0259 0261(154) (155) (162) (163) (153) (154) (161) (162)

LOGTNAit1 0058 0058 0063 0061 0055 0056 0060 0058(315) (317) (340) (329) (303) (305) (328) (317)


LOGFAMSIZEit1 0012 0012 0011 0011 0012 0012 0012 0012(257) (257) (254) (254) (259) (259) (257) (256)

TURNOVERit1 0000 0000 0000 0000 0000 0000 0000 0000(162) (161) (162) (161) (157) (156) (157) (157)

AGEit1 0001 0001 0001 0001 0001 0001 0001 0001(098) (097) (098) (097) (087) (087) (088) (087)

EXPRATIOit1 0018 0018 0019 0019 0062 0062 0064 0063(044) (043) (046) (044) (149) (148) (152) (151)

TOTLOADit1 0001 0001 0001 0001 0001 0001 0001 0001(054) (054) (052) (052) (048) (047) (046) (046)

FLOWit1 0000 0000 0000 0000 0000 0000 0000 0000(096) (095) (090) (092) (102) (102) (096) (099)

FUNDRETit1 0022 0022 0021 0021 0022 0022 0022 0022(345) (345) (343) (342) (348) (348) (345) (345)

No of months 228 228 228 228 228 228 228 228




INTERCEPT 0016 0068 0128 0068 0017 0035 0095 0034(019) (084) (172) (092) (020) (043) (127) (047)

INDLC 0119 0121 0119 0120 0112 0114 0112 0113(131) (133) (133) (134) (123) (125) (124) (126)

LOGTNAit1 0029 0029 0032 0029 0027 0027 0029 0027(222) (224) (245) (225) (202) (205) (226) (205)

LOGTNAit1 0031 0031 0031 0031 0030 0030 0030 0030INDLC (229) (232) (234) (236) (223) (226) (228) (230)

LOGFAMSIZEit1 0011 0011 0011 0011 0011 0011 0011 0011(254) (254) (251) (251) (258) (258) (255) (255)

TURNOVERit1 0000 0000 0000 0000 0000 0000 0000 0000(180) (179) (180) (179) (175) (175) (176) (175)

AGEit1 0001 0001 0001 0001 0001 0001 0001 0001(126) (125) (126) (125) (117) (117) (117) (116)

EXPRATIOit1 0020 0019 0021 0020 0065 0065 0066 0065(049) (048) (052) (050) (163) (162) (166) (164)

TOTLOADit1 0001 0001 0001 0001 0001 0001 0001 0001(035) (035) (034) (035) (030) (030) (029) (029)

FLOWit1 0000 0000 0000 0000 0000 0000 0000 0000(084) (082) (077) (079) (089) (088) (083) (085)

FUNDRETit1 0020 0020 0020 0020 0021 0020 0020 0020(310) (309) (307) (307) (313) (312) (310) (309)

No of months 228 228 228 228 228 228 228 228












LOGTNAit1 0051 0052 0056 0054 0048 0049 0053 0051(283) (286) (313) (302) (270) (272) (301) (289)


LOGFAMSIZEit1 0007 0007 0007 0007 0007 0007 0007 0007(070) (070) (069) (070) (070) (070) (069) (070)


No of months 228 228 228 228 228 228 228 228




LOGTNAit1 0067 0070 0071 0072 0064 0067 0068 0069(367) (383) (387) (391) (351) (367) (371) (376)


LOGFAMSIZEit1 0012 0012 0012 0012 0012 0012 0012 0012(113) (113) (111) (111) (115) (115) (113) (113)


No of months 228 228 228 228 228 228 228 228




































(1) (2) (3)

SMALLFUND 00759 00580 00584(949) (674) (679)

SMALLCAPFUND 00515 00340 00341(628) (386) (388)














(1) (2) (3)

SMALLFUND 00258 00143 00150(183) (094) (099)

SMALLCAPFUND 01190 01078 01079(821) (695) (696)














STOCKS


MULTI-MANAGER 00102 00190(175) (173)


















V Conclusion





INTERCEPT 0069 0018 0039 0008 0092 0042 0014 0032(066) (019) (025) (005) (088) (044) (009) (021)

MULTI-MANAGERi 0040 0040 0041 0041 0040 0041 0042 0042(222) (223) (230) (230) (226) (227) (234) (234)

LOGTNAit1 0024 0023 0026 0025 0022 0022 0025 0023(234) (230) (198) (183) (219) (215) (186) (170)

LOGFAMSIZEit1 0016 0016 0016 0016 0016 0016 0016 0016(238) (238) (238) (238) (240) (239) (240) (239)

TURNOVERit1 0000 0000 0000 0000 0000 0000 0000 0000(128) (129) (128) (129) (126) (126) (126) (126)

AGEit1 0001 0001 0001 0001 0001 0001 0001 0001(064) (065) (064) (064) (060) (061) (059) (059)

EXPRATIOit1 0002 0002 0003 0003 0043 0043 0045 0045(003) (003) (004) (004) (054) (054) (055) (055)

TOTLOADit1 0002 0002 0002 0002 0002 0002 0002 0002(028) (029) (030) (029) (036) (036) (037) (037)

FLOWit1 0000 0000 0000 0000 0000 0000 0000 0000(164) (168) (168) (169) (164) (168) (168) (169)

FUNDRETit1 0030 0030 0029 0029 0031 0030 0030 0030(338) (338) (334) (334) (341) (341) (337) (337)

No of months 96 96 96 96 96 96 96 96





REFERENCES





































































LOGTNAit1 0051 0052 0056 0054 0048 0049 0053 0051(283) (286) (313) (302) (270) (272) (301) (289)


LOGFAMSIZEit1 0007 0007 0007 0007 0007 0007 0007 0007(070) (070) (069) (070) (070) (070) (069) (070)


No of months 228 228 228 228 228 228 228 228




LOGTNAit1 0067 0070 0071 0072 0064 0067 0068 0069(367) (383) (387) (391) (351) (367) (371) (376)


LOGFAMSIZEit1 0012 0012 0012 0012 0012 0012 0012 0012(113) (113) (111) (111) (115) (115) (113) (113)


No of months 228 228 228 228 228 228 228 228




































(1) (2) (3)

SMALLFUND 00759 00580 00584(949) (674) (679)

SMALLCAPFUND 00515 00340 00341(628) (386) (388)














(1) (2) (3)

SMALLFUND 00258 00143 00150(183) (094) (099)

SMALLCAPFUND 01190 01078 01079(821) (695) (696)














STOCKS


MULTI-MANAGER 00102 00190(175) (173)


















V Conclusion





INTERCEPT 0069 0018 0039 0008 0092 0042 0014 0032(066) (019) (025) (005) (088) (044) (009) (021)

MULTI-MANAGERi 0040 0040 0041 0041 0040 0041 0042 0042(222) (223) (230) (230) (226) (227) (234) (234)

LOGTNAit1 0024 0023 0026 0025 0022 0022 0025 0023(234) (230) (198) (183) (219) (215) (186) (170)

LOGFAMSIZEit1 0016 0016 0016 0016 0016 0016 0016 0016(238) (238) (238) (238) (240) (239) (240) (239)

TURNOVERit1 0000 0000 0000 0000 0000 0000 0000 0000(128) (129) (128) (129) (126) (126) (126) (126)

AGEit1 0001 0001 0001 0001 0001 0001 0001 0001(064) (065) (064) (064) (060) (061) (059) (059)

EXPRATIOit1 0002 0002 0003 0003 0043 0043 0045 0045(003) (003) (004) (004) (054) (054) (055) (055)

TOTLOADit1 0002 0002 0002 0002 0002 0002 0002 0002(028) (029) (030) (029) (036) (036) (037) (037)

FLOWit1 0000 0000 0000 0000 0000 0000 0000 0000(164) (168) (168) (169) (164) (168) (168) (169)

FUNDRETit1 0030 0030 0029 0029 0031 0030 0030 0030(338) (338) (334) (334) (341) (341) (337) (337)

No of months 96 96 96 96 96 96 96 96





REFERENCES





























































































(1) (2) (3)

SMALLFUND 00759 00580 00584(949) (674) (679)

SMALLCAPFUND 00515 00340 00341(628) (386) (388)














(1) (2) (3)

SMALLFUND 00258 00143 00150(183) (094) (099)

SMALLCAPFUND 01190 01078 01079(821) (695) (696)














STOCKS


MULTI-MANAGER 00102 00190(175) (173)


















V Conclusion





INTERCEPT 0069 0018 0039 0008 0092 0042 0014 0032(066) (019) (025) (005) (088) (044) (009) (021)

MULTI-MANAGERi 0040 0040 0041 0041 0040 0041 0042 0042(222) (223) (230) (230) (226) (227) (234) (234)

LOGTNAit1 0024 0023 0026 0025 0022 0022 0025 0023(234) (230) (198) (183) (219) (215) (186) (170)

LOGFAMSIZEit1 0016 0016 0016 0016 0016 0016 0016 0016(238) (238) (238) (238) (240) (239) (240) (239)

TURNOVERit1 0000 0000 0000 0000 0000 0000 0000 0000(128) (129) (128) (129) (126) (126) (126) (126)

AGEit1 0001 0001 0001 0001 0001 0001 0001 0001(064) (065) (064) (064) (060) (061) (059) (059)

EXPRATIOit1 0002 0002 0003 0003 0043 0043 0045 0045(003) (003) (004) (004) (054) (054) (055) (055)

TOTLOADit1 0002 0002 0002 0002 0002 0002 0002 0002(028) (029) (030) (029) (036) (036) (037) (037)

FLOWit1 0000 0000 0000 0000 0000 0000 0000 0000(164) (168) (168) (169) (164) (168) (168) (169)

FUNDRETit1 0030 0030 0029 0029 0031 0030 0030 0030(338) (338) (334) (334) (341) (341) (337) (337)

No of months 96 96 96 96 96 96 96 96





REFERENCES




































































(1) (2) (3)

SMALLFUND 00258 00143 00150(183) (094) (099)

SMALLCAPFUND 01190 01078 01079(821) (695) (696)














STOCKS


MULTI-MANAGER 00102 00190(175) (173)


















V Conclusion





INTERCEPT 0069 0018 0039 0008 0092 0042 0014 0032(066) (019) (025) (005) (088) (044) (009) (021)

MULTI-MANAGERi 0040 0040 0041 0041 0040 0041 0042 0042(222) (223) (230) (230) (226) (227) (234) (234)

LOGTNAit1 0024 0023 0026 0025 0022 0022 0025 0023(234) (230) (198) (183) (219) (215) (186) (170)

LOGFAMSIZEit1 0016 0016 0016 0016 0016 0016 0016 0016(238) (238) (238) (238) (240) (239) (240) (239)

TURNOVERit1 0000 0000 0000 0000 0000 0000 0000 0000(128) (129) (128) (129) (126) (126) (126) (126)

AGEit1 0001 0001 0001 0001 0001 0001 0001 0001(064) (065) (064) (064) (060) (061) (059) (059)

EXPRATIOit1 0002 0002 0003 0003 0043 0043 0045 0045(003) (003) (004) (004) (054) (054) (055) (055)

TOTLOADit1 0002 0002 0002 0002 0002 0002 0002 0002(028) (029) (030) (029) (036) (036) (037) (037)

FLOWit1 0000 0000 0000 0000 0000 0000 0000 0000(164) (168) (168) (169) (164) (168) (168) (169)

FUNDRETit1 0030 0030 0029 0029 0031 0030 0030 0030(338) (338) (334) (334) (341) (341) (337) (337)

No of months 96 96 96 96 96 96 96 96





REFERENCES




































































STOCKS


MULTI-MANAGER 00102 00190(175) (173)


















V Conclusion





INTERCEPT 0069 0018 0039 0008 0092 0042 0014 0032(066) (019) (025) (005) (088) (044) (009) (021)

MULTI-MANAGERi 0040 0040 0041 0041 0040 0041 0042 0042(222) (223) (230) (230) (226) (227) (234) (234)

LOGTNAit1 0024 0023 0026 0025 0022 0022 0025 0023(234) (230) (198) (183) (219) (215) (186) (170)

LOGFAMSIZEit1 0016 0016 0016 0016 0016 0016 0016 0016(238) (238) (238) (238) (240) (239) (240) (239)

TURNOVERit1 0000 0000 0000 0000 0000 0000 0000 0000(128) (129) (128) (129) (126) (126) (126) (126)

AGEit1 0001 0001 0001 0001 0001 0001 0001 0001(064) (065) (064) (064) (060) (061) (059) (059)

EXPRATIOit1 0002 0002 0003 0003 0043 0043 0045 0045(003) (003) (004) (004) (054) (054) (055) (055)

TOTLOADit1 0002 0002 0002 0002 0002 0002 0002 0002(028) (029) (030) (029) (036) (036) (037) (037)

FLOWit1 0000 0000 0000 0000 0000 0000 0000 0000(164) (168) (168) (169) (164) (168) (168) (169)

FUNDRETit1 0030 0030 0029 0029 0031 0030 0030 0030(338) (338) (334) (334) (341) (341) (337) (337)

No of months 96 96 96 96 96 96 96 96





REFERENCES










































































V Conclusion





INTERCEPT 0069 0018 0039 0008 0092 0042 0014 0032(066) (019) (025) (005) (088) (044) (009) (021)

MULTI-MANAGERi 0040 0040 0041 0041 0040 0041 0042 0042(222) (223) (230) (230) (226) (227) (234) (234)

LOGTNAit1 0024 0023 0026 0025 0022 0022 0025 0023(234) (230) (198) (183) (219) (215) (186) (170)

LOGFAMSIZEit1 0016 0016 0016 0016 0016 0016 0016 0016(238) (238) (238) (238) (240) (239) (240) (239)

TURNOVERit1 0000 0000 0000 0000 0000 0000 0000 0000(128) (129) (128) (129) (126) (126) (126) (126)

AGEit1 0001 0001 0001 0001 0001 0001 0001 0001(064) (065) (064) (064) (060) (061) (059) (059)

EXPRATIOit1 0002 0002 0003 0003 0043 0043 0045 0045(003) (003) (004) (004) (054) (054) (055) (055)

TOTLOADit1 0002 0002 0002 0002 0002 0002 0002 0002(028) (029) (030) (029) (036) (036) (037) (037)

FLOWit1 0000 0000 0000 0000 0000 0000 0000 0000(164) (168) (168) (169) (164) (168) (168) (169)

FUNDRETit1 0030 0030 0029 0029 0031 0030 0030 0030(338) (338) (334) (334) (341) (341) (337) (337)

No of months 96 96 96 96 96 96 96 96





REFERENCES






























































REFERENCES




























































does fund size erode mutual fund performance the role of liquidity and organization(3)

Economy & Finance