Example: biology

Does Asset Allocation Policy Explain 40, 90, or 100 ...

26 2000, Association for Investment Management and ResearchDoes Asset Allocation Policy Explain 40, 90, or 100 Percent of Performance?Roger G. Ibbotson and Paul D. KaplanDisagreement over the importance of Asset Allocation Policy stems fromasking different questions. We used balanced mutual fund and pensionfund data to answer the three relevant questions. We found that about 90percent of the variability in returns of a typical fund across time is explainedby Policy , about 40 percent of the variation of returns among funds isexplained by Policy , and on average about 100 percent of the return level isexplained by the Policy return Asset Allocation Policy Explain 40 per-cent, 90 percent, or 100 percent of perfor-mance? The answer depends on how thequestion is asked and what an analyst istrying to Explain .

Financial Analysts Journal 28 ©2000, Association for Investment Management and Research Stevens, Surz, and Wimer (1999) provided the same type of analysis on quarterly returns of 58 pension funds over the five-year 1993–97 period.4 We used the actual policy weights and asset-class benchmarks of the pension funds, however, rather

Tags:

  Return

Information

Domain:

Source:

Link to this page:

Please notify us if you found a problem with this document:

Other abuse

Transcription of Does Asset Allocation Policy Explain 40, 90, or 100 ...

1 26 2000, Association for Investment Management and ResearchDoes Asset Allocation Policy Explain 40, 90, or 100 Percent of Performance?Roger G. Ibbotson and Paul D. KaplanDisagreement over the importance of Asset Allocation Policy stems fromasking different questions. We used balanced mutual fund and pensionfund data to answer the three relevant questions. We found that about 90percent of the variability in returns of a typical fund across time is explainedby Policy , about 40 percent of the variation of returns among funds isexplained by Policy , and on average about 100 percent of the return level isexplained by the Policy return Asset Allocation Policy Explain 40 per-cent, 90 percent, or 100 percent of perfor-mance? The answer depends on how thequestion is asked and what an analyst istrying to Explain .

2 According to well-known studiesby Brinson and colleagues, more than 90 percent ofthe variability in a typical plan sponsor s perfor-mance over time is the result of Asset So, if one is trying to Explain the variabilityof returns over time, Asset Allocation is very , the Brinson et al. studies areoften misinterpreted and the results applied toquestions that the studies never intended toanswer. For example, an analyst might want toknow how important Asset Allocation is in Explain -ing the variation of performance among the Brinson studies did not address thisquestion, the analyst can neither look to them tofind the answer nor fault them for not answering A different study is , an analyst might want to know whatpercentage of the level of a typical fund s return isascribable to Asset Allocation Policy .

3 Again, theBrinson studies do not address this question. Adifferent study is needed. Thus, three distinct questions remain about theimportance of Asset much of the variability of returns acrosstime is explained by Policy (the question Brin-son et al. asked)? In other words, how much ofa fund s ups and downs do its Policy bench-marks Explain ? much of the variation in returns amongfunds is explained by differences in Policy ? Inother words, how much of the differencebetween two funds performance is a result oftheir Policy difference? portion of the return level is explained bypolicy return ? In other words, what is the ratioof the Policy benchmark return to the fund sactual return ?Much of the recent controversy about theimportance of Asset Allocation stems from treatingthe answer that Brinson et al.

4 Provided to Question1 as an answer to Questions 2 and purpose of our study was to ask andanswer all three questions. To do this, we examined10 years of monthly returns to 94 balancedmutual funds and 5 years of quarterly returns to 58pension funds. We performed a different analysisfor each data consisted of the total return for each fundfor each period of time (a month or a quarter). Thefirst step in our analysis was to decompose eachtotal return , TR, into two components, Policy returnand active return , as follows:TRi,t = (1 + PRi,t)(1 + ARi,t) 1,whereTRi,t= total return of fund i in period tPRi,t= Policy return of fund i in period tARi,t= active return of fund i in period tRoger G. Ibbotson is professor of finance at the YaleSchool of Management and chair of Ibbotson D.

5 Kaplan is the director of the Morningstar Centerfor Quantitative Research. He was vice president andchief economist at Ibbotson Associates when this articlewas 2000, Financial Analysts and republished with permissionfrom the Association for Investment Manage-ment and Research. All rights Asset Allocation Policy Explain 40, 90, or 100 Percent of Performance?January/February 200027 Policy return is the part of the total return thatcomes from the Asset Allocation Policy . Activereturn is the remainder. Active return depends onboth the manager s ability to actively over- orunderweight Asset classes and securities relative tothe Policy and on the magnitude and timing ofthose Asset Allocation Policy of each fund can berepresented as a set of Asset -class weights that sumto 1.

6 For the pension funds in this study, theseweights were known in advance. For the mutualfunds, the Policy weights were determined byreturn-based style analysis, which is described inthe Data section. The Policy return of the fundover a given period of time can be computed fromthe Policy weights and returns on Asset -classbenchmarks as follows:PRi,t = w1iR1t + w2iR2t + .. + wkiRkt c,wherew1i, w2i, .., wki= Policy weights of fund iR1t, R2t, .., Rkt= returns on the Asset classesin period tc = approximate cost of replicat-ing the Policy mix throughindexed mutual funds, as apercentage of assetsThus, in addition to fund returns, we needed policyweights for each fund and total returns onasset-class benchmarks. Given the total returns tothe funds and the estimated Policy returns of thefunds, we solved for the active our time-series analysis, we used theperiod-by-period returns.

7 In our cross-sectionalanalysis, we used the compound annual rates ofreturn over the period of analysis. For each fund,we computed the compound annual total returnover the entire period as follows:whereTRi=compound annual total return onfund i over the entire period ofanalysisTRi,t=total return of fund i in period tT=number of period returnsN=length of the entire period ofanalysis, in yearsSimilarly, we computed the compound annualpolicy return over the entire period as follows:where PRi is the compound annual Policy return onfund i over the entire period of analysis and PRi,t isthe Policy return to fund i in period the mutual fund portion of this study, we used10 years of monthly returns for 94 balancedfunds. The 94 funds represent all of the balancedfunds in the Morningstar universe that had at least10 years of data ending March 31, 1998.

8 Policyweights for each fund were estimated by perform-ing return -based style analysis over the entire120-month Ta b l e 1 shows the Asset -classbenchmarks used and the average fund exposureto each Asset calculating the Policy returns for each fund,we assumed that the cost of replicating the policymix through index mutual funds would be 2 basispoints a month (approximately 25 bps annually).TRi1 TRi1,+()1 TRi2,+()..1 TRiT,+()N1, =PRi1 PRi1,+()1 PRi2,+()..1 PRiT,+()N1, =Table 1. Asset Classes and Benchmarks for Balanced Mutual FundsAsset ClassBenchmarkAverage AllocationLarge-cap stocksCRSP 1 2 stocksCRSP 6 8 stocksMSCI Europe/Australasia/Far East bondsLehman Brothers Aggregate Bond by CRSP. CRSP excludes unit investment trusts, closed-end funds, real estate investmenttrusts, Americus trusts, foreign stocks, and American Depositary Receipts from the portfolios.

9 CRSP usesonly NYSE firms to determine the size breakpoints for the portfolios. Specifically, CRSP ranks all eligibleNYSE stocks by company size (market value of outstanding equity) and then splits them into 10 equallypopulated groups, or deciles. The largest companies are in Decile 1, and the smallest are in Decile 10. Thecapitalization for the largest company in each decile serves as the breakpoint for that decile. Breakpointsare rebalanced on the last day of trading in March, June, September, and December. CRSP then assignsNYSE and Amex/Nasdaq companies to the portfolios according to the decile breakpoints. Monthlyportfolio returns are market-cap-weighted averages of the individual returns within each of the 10portfolios. The 1 2 portfolio is the combination of Deciles 1 and 2, and the 6 8 portfolio is the combinationof Deciles 6, 7, and Associates (1998).

10 Financial Analysts Journal28 2000, Association for Investment Management and ResearchStevens, Surz, and Wimer (1999) provided thesame type of analysis on quarterly returns of 58pension funds over the five-year 1993 97 used the actual Policy weights and Asset -classbenchmarks of the pension funds, however, ratherthan estimated Policy weights and the sameasset-class benchmarks for all funds. In each quar-ter, the Policy weights were known in advance ofthe realized We report the pension fundresults together with our analysis of the mutualfund returns in the next and AnswersNow consider the original three questions posed bythe study: How much of the variability of returnacross time is explained by Asset Allocation Policy ,how much of the variation among funds isexplained by the Policy , and what portion of thereturn level is explained by Policy return ?


Related search queries