FINSER_23_3 Using the new portability election of deceased spouses: A pedagogical example De’Arno De’Armond, Ph.D.a,*, Darlene Pulliam, Ph.D.b, Robin Patterson, J.D.c aEdwards Professor of Financial Planning, West Texas A&M University, Box 60809, Canyon, TX 79016-001, USA bRegents Professor of Accounting, McCray Professor of Business, West Texas A&M University, Box 60809, Canyon, TX 79016-001, USA cPatterson Professor of Business Law, West Texas A&M University, Box 60809, Canyon, TX 79106-001, USA Abstract The United States has a unified system that taxes transfers of property during an individual’s lifetime (gifts) and property transferred as a result of the individual’s death. The Tax Relief, Unemployment Insurance Reauthorization, and Job Creation Act of 2010 (the Act) contains a provision that will allow the unused portion of a decedent’s exclusion (taxable estate protected by the unified credit) to be used upon the subsequent death of the surviving spouse. The portability election is simple for situations where it appears the surviving spouse will not remarry, however, becomes much more complicated if the surviving spouse should remarry. © 2014 Academy of Financial Services. All rights reserved. JEL classification: M410, Accounting; K11, Estate Planning Keywords: Tax accounting; Estate planning; Financial planning; Lifetime gifts; Portability election 1. Introduction On the minds of many financial planners, advisors, practitioners, estate planners, tax accountants, academics, and general individuals as well, resides pending implications and effects of the “Fiscal Cliff” facing Americans (Dudley, 2013). Finding fiscal balance between political parties may be challenging for the United States at best (Grunwald, 2012). Whereas * Corresponding author. Tel.: !1-806-651-2490; fax: !1-806-651-2488. E-mail address: ddearmond@wtamu.edu (D. De’Armond) Financial Services Review 23 (2014) 239–248 1057-0810/14/$ – see front matter © 2014 Academy of Financial Services. All rights reserved. many facets of the “Cliff” are known and commonly explored, there exist many smaller facets such as a “slew of temporary tax policies” that have, for the most part, been underexposed with exception to discussions often happening within the confines of the academic classroom and or certain individual prevailing circumstances (Suderman, 2013). As political agreements have surfaced during this time, estate planning for married couples has been impacted, as evidenced with the passing of the Tax Relief Act of 2010 (Deener, 2012). The passing of the “the Act” brought with it one such facet, dubbed DSUEA, the “deceased spousal unused exclusion amount.” DSUEA provides a surviving spouse an option to utilize the unused applicable exclusion of his or her deceased spouse for decedents dying after 2010 (Dunn, 2011). This transfer or “portability” election requires special care and consideration in its application and should not be seen as a “cure-all” as many potential complex issues may arise (Gallo, 2011). Careful attention should be given the caveats of the portability election in an effort to utilize portability in the most efficient and effective manner (Katz, 2011). As one might expect, the Internal Revenue Service issued bulletins appropriate to the filings of DSUEA via form 706 as a vehicle (Internal Revenue Bulletin, 2011). Although information has surfaced cautioning the use of the portability election, it remains a viable option given certain estate conditions (Joseph, 2012). Most recently, with the passing of the American Taxpayer Relief Act of 2012 (ATRA), congress made permanent the two signif- icant provisions discussed within this article, the $5,000,000 indexed basic exclusion amount and DSUEA, as well as introduced a maximum transfer tax rate of 40% (Miller, 2013). As other authors and academics alike have indicated in their writings, the public tends to not possess the financial knowledge to make good financial decisions all of the time (Beierlein and Neverett, 2013). It is in the spirit of encouraging basic and applied research in the area of personal financial planning that this work is presented. This article serves to enlighten and inform readers on one such “portability of deceased spouse unused exclusion tax provision” providing a pedagogical example utilized within the university setting in an effort to disseminate information and assist the development of curricula in financial services. In much the same “learning by doing” spirit that Thomas Eyssell wrote about the financial planning process, this particular work contributes to the literature (Eyssell, 1999). This particular work will serve as a quick review of the U.S. estate and gift law, with brief discussions of the gift tax, annual exclusion, marital deduction, generation skipping tax, tax planning credits, and the new portability elections. The work concludes with examples and an Excel application model developed for use in the academic and/or learning environment, and summary of the possible implications. 2. A quick review of U.S. estate and gift law The United States has a unified system that taxes transfers of property during an individ- ual’s lifetime (gifts) and property transferred as a result of the individual’s death. The system is structured so that the sum of an individual’s taxable gifts and net assets transferred at death must exceed an applicable exemption amount before any gift or estate taxes are imposed. The 240 D. De’Armond et al. / Financial Services Review 23 (2014) 239–248 applicable exemption amount is $5,000,000 for 2011, $5,120,000 in 2012, $5,250,000 in 2013, and $5,340,000 in 2014. Table 1, Unified Transfer Tax Rate Schedule, After 2012 provides information for the sum of taxable estate and adjusted taxable gifts and their applicable tentative tax. As an example, utilizing Table 1 below, the tentative tax on $5,340,000 using Unified Transfer Tax Rate Schedule is found as $2,081,800 less applicable Unified Credit for 2014 of $2,081,800, yielding a net transfer tax payable of $0. Note that in 2014 $5,340,000 of net estate is exempted from tax, but the marginal tax rate on the first dollar over $5,340,000 is 40%. 2.1. The gift tax The gift tax is imposed on a calendar year basis on transfers of property by gift by any individual. Internal Revenue Code Section (Sec.) 2501. It is cumulative during an individ- ual’s life. The gift tax is the excess of a tentative tax, computed on the aggregate sum of the taxable gifts for the calendar year and the preceding calendar years, over a tentative tax computed on the aggregate sum of the taxable gifts for the preceding calendar years. Sec. 2502(a). The tax is computed with the same tax rates used for estate purposes in Sec. 2001(c). 2.2. The annual exclusion In defining the term “taxable gifts,” Sec. 2503 allows for an annual exclusion of $14,000 (in 2014) for gifts to any number of persons each year. Since 1998, the amount of the annual exclusion has been indexed annually for inflation in increments of $1,000. Gifts of a future interest do not generally qualify for the exclusion. However, transfers for the benefit of individuals under the age of 21 will not be considered transfers of a future interest if the property and the income therefrom may be expended by, or for the benefit of the donee before attainment of the age of 21 and any amounts not so expended will pass to the donee upon that donee’s 21st birthday. Sec. 2503(c). Table 1 Unified transfer tax rate schedule after 2012 Sum of taxable estate and adjusted taxable gifts Tentative tax Over But not under $0 $ 10,000 18% of such amount $10,000 $ 20,000 $1,800 plus 20% of excess over $10,000 $20,000 $ 40,000 $3,800 plus 22% of excess over $20,000 $40,000 $ 60,000 $8,200 plus 24% of excess over $40,000 $60,000 $ 80,000 $13,000 plus 26% of excess over $60,000 $80,000 $ 100,000 $18,200 plus 28% of excess over $80,000 $100,000 $ 150,000 $23,800 plus 30% of excess over $100,000 $150,000 $ 250,000 $38,800 plus 32% of excess over $150,000 $250,000 $ 500,000 $70,800 plus 34% of excess over $250,000 $500,000 $ 750,000 $155,800, plus 37% of the excess over $500,000 $750,000 $1 million $248,300, plus 39% of the excess over $750,000 $1 million $345,800, plus 40% of the excess over $1 million 241D. De’Armond et al. / Financial Services Review 23 (2014) 239–248 This exclusion allows for the transfer of substantial property over a period of years without the imposition of either the gift or estate tax. For example, an individual with five children can transfer $70,000 (5 " $14,000) per year to his or her children. A couple electing gift-splitting under Sec. 2513 can transfer $140,000 per year to five children with no gift or estate tax liability. These amounts should increase because of the annual exclusion inflation adjustments for gifts made after 1998. A gift-giving program can result in the transfer of large amounts of property over several years with no gift or estate tax implications. 2.3. The marital deduction A donor spouse is allowed an unlimited deduction for lifetime gifts made to his or her spouse (Sec. 2523). This allows for gifts between spouses with no gift or estate tax implications. Certain terminable interests, interests that will terminate after a certain period or on the occurrence of some event, do not qualify for the marital deduction (Sec. 2523). However, if it meets the requirements of Sec. 2523(f), a life estate may qualify for the marital deduction if it is “qualified terminable interest property.” Property is qualified terminable interest property (QTIP) if: Y The donee spouse is entitled to the income from the property for life; Y Income from the property is payable to the donee spouse at least annually; Y No person has the power to appoint any part of the property to any person other than the donee spouse; and Y An irrevocable election to have all or part of the trust qualify for the marital deduction is made by the donor and attached to the gift tax return (Sec. 2523). 2.4. The generation skipping tax To remove the benefit of transferring property to grandchildren or otherwise skipping the estate or gift tax that will be paid by the immediately after generation, a generation-skipping transfer tax is imposed at a flat 40% rate on generation-skipping transfers. The most common generation-skipping transfers are direct transfers from a grandparent to a grandchild and life estates from the grandparent to the child with the remainder interest to the grandchild. Generation-skipping transfers are defined in Sec. 2603. These transfers include taxable distributions, taxable terminations, and direct skips. A taxable distribution is a distribution to a transferee who is a member of a generation at least two generations younger than the transferor. The amount received is subject to the tax (Sec. 2611). The tax is paid by the transferee (Sec. 2603). A taxable termination is the termination by death, lapse of time, release of power, or similar event of an interest held in trust that passes to a transferee who is a member of a generation at least two generations younger than the transferor. The value of the property in which the interest terminates is subject to the tax (Sec. 2611). The tax is paid by the trustee (Sec. 2603). A direct skip is a transfer of property to, or for the benefit of persons two or more generations below the transferor. The value of the property transferred is subject to the tax (Sec. 2611). The tax is paid by the transferor (Sec. 2603). There is a lifetime generation-skipping tax exemption of 242 D. De’Armond et al. / Financial Services Review 23 (2014) 239–248 $5,340,000 (in 2014) per grantor (Sec. 2631). This amount will be indexed for inflation for decedents dying and gifts made after 2011 (Code Sec. 2631(c)). 2.5. Couples or individuals with $5,340,000 or less. Because of the structure of the unified transfer system, couples with net assets of $5,340,000 or less (smaller applicable unified credit for years earlier than 2014) will pay no federal estate tax. The only transfer tax planning that might need to be done would relate to state estate or inheritance taxes. The primary assistance that the tax advisor can provide is a strong recommendation that the individual execute a will, a durable power of attorney and an advance directive to physician. A revocable trust might also be considered. Even with small estates, nontax considerations might warrant estate planning to protect the surviving spouse and other family members. There are some situations where one or both of the spouses might not want his or her assets to pass directly to the surviving spouse. One of the spouses might not be capable of managing the couple’s assets. There could be concern that the surviving spouse’s remarriage might result in the original couple’s assets coming under the control of the new spouse, and consequently, not being available for the offspring of the first marriage. In estate planning for a second marriage, one or both of the parties might want to provide for the surviving spouse during his or her lifetime, but want the assets to be available for the offspring of an earlier marriage upon the death of the surviving spouse. In these cases, creation of a trust might be in order. One option would be the creation of a revocable trust with a provision that all or part of the trust becomes irrevocable upon the death of one of the spouses. Another option would be a testamentary trust created under the will of one or both of the spouses. In either case, the trust could provide for the income of the assets and some limited portion of the assets to be available to the surviving spouse during his or her lifetime, with the remaining assets to be distributed to their beneficiaries upon that spouse’s death. 3. Former tax planning to use the unified credit of both spouses Before the Act discussed below, the most basic estate planning technique was to use both spouses’ unified credit by using the marital deduction effectively. This was not accomplished if all of the assets of the first spouse to die pass freely to the surviving spouse. In this case, the first to die has no net estate because of the marital deduction and no need for the unified credit. However, unless the surviving spouse manages to spend enough of the combined assets to reduce his or her estate to less than $5,340,000, this second spouse’s unified credit will not be enough to shield that estate from estate taxes. One approach was to leave all of the assets except $5,340,000 (or, if less, enough assets to reduce the second estate to $5,340,000) to the surviving spouse. The marital deduction would reduce the first estate to $5,340,000 and the first unified credit would offset all of the taxes. The second unified credit could then offset up to $5,340,000 of the second estate, with maximum use being made of both unified credits. Except for very large estates, a direct transfer of assets to someone other than the surviving 243D. De’Armond et al. / Financial Services Review 23 (2014) 239–248 spouse might not be acceptable because the surviving spouse needs the assets, or at least the income from the assets, for living expenses. However, the use of a trust with a terminable interest that does not qualify as qualified terminable interest property could yield the same results. An effective, flexible tool is a qualified terminable interest property trust giving the executor the right to make the election as to how much of the life estate will be treated as qualified terminable interest property qualifying for the marital deduction. The executor will choose to not make the election for the amount of the property equal to the applicable exemption amount. A proper election will result in just enough property being left in the taxable estate to use up the decedent’s applicable unified credit. If a couple owned all of its property jointly, this technique could not be used successfully. Jointly owned property passes automatically to the surviving owner. This property would qualify for the marital deduction, leaving nothing in the estate of the first spouse to die and everything in the estate of the surviving spouse. Marital deduction planning could not be achieved if one of the spouses owned none or very little of the assets of the couple and that spouse dies first. There was no way for that spouse to retain assets to use up his or her unified credit. To facilitate this planning technique, the spouse owning the assets could make appropriate gifts to the other spouse. The unlimited marital deduction would prevent any taxable gifts. Trusts serve estate planning purposes other than marital deduction planning. A trust with the income to be paid to the surviving spouse and the remaining assets to be paid to the couple’s children or other named beneficiaries upon the second death can be used to protect the assets for the benefit of both the surviving spouse and the beneficiaries. Putting the assets in trust keeps the assets out of the control and out of the estate of a subsequent spouse of the decedent’s surviving spouse. If the trusts are set up so that they would qualify as qualified terminable interest property if the executor so elects, the executor of the estate can make the proper election to result in a marital deduction that maximizes the use of the couple’s applicable unified credits. 4. Availability of new portability elections The Tax Relief, Unemployment Insurance Reauthorization, and Job Creation Act of 2010 (the Act) contains a provision that will allow the unused portion of a decedent’s exclusion (taxable estate protected by the unified credit) to be used upon the subsequent death of the surviving spouse. The American Taxpayer Relief Act of 2012 extended the portability election indefinitely. The Act fixed the amount of taxable estate protected by the exclusion—the basic exclusion amount—at $5,000,000 through December 31, 2011. This amount has been indexed annually for decedents dying after 2011—resulting in $5,120,000 for 2012, $5,250,000 in 2013 and $5,340,000 in 2014. Under IRC Sec. 2010(c)(2) the applicable exclusion amount is the sum of Y The basic exclusion amount and, 244 D. De’Armond et al. / Financial Services Review 23 (2014) 239–248 Y If a proper election has been made for a surviving spouse, the deceased spouse’s unused exclusion amount. Under IRC Sec. 2010(c)(4) the deceased spouse’s unused exclusion amount is the lesser of: Y The basic exclusion amount or Y The excess of ! The basic exclusions amount of the last deceased spouse of the surviving spouse over, ! The tentative tax computed on the estate of that deceased spouse. The Act goes on to require that the executor of the estate of the deceased must file an estate tax return computing the tax and making an election to make the unused exclusion available to the surviving spouse’s estate or the deceased spouse unused exclusion cannot be taken into account. The Act also allows that the returns for which a portability election has been made can be examined after the statute of limitations has expired. The deceased spousal unused exclusion can be reduced or eliminated with this examination. However, no additional taxes can be assessed on the original return after the statute has expired. 4.1. If surviving spouse is not expected to remarry The portability election is fairly simple for situations where it appears the surviving spouse will not remarry. If all of the decedent’s assets pass to the surviving spouse, the marital deduction will reduce the taxable estate to zero and no exclusion will be needed to reduce the taxes due. Bequests to children or other parties will reduce the marital deduction and increase the taxable estate—using some or all of the exclusion. If any of the exclusion is not used, the executor should make the election on the 706 when it is filed. 4.2. Example Mr. and Mrs. Smith have net assets of $7,500,000, owned equally by each spouse. Mr. Smith dies in 2014, leaving all of his assets directly to Mrs. Smith. Mr. Smith’s estate will have no net estate because of the marital deduction. The executor of Mr. Smith’s estate should make the election to make the unused exclusion available to Mrs. Smith’s estate. If Mrs. Smith still has $7,500,000 at death, her estate will use the combined exclusions from both estates to offset of the tentative tax on her estate. 4.3. Surviving spouse is expected to remarry In the past, in marriages where the spouse died first, employment history of the couple, property laws and estate planning often resulted in the spouse owning more assets than the spouse. Consequently, if the spouse died first owning assets equaling less than the exclusion amount less than the total exclusion was used. Although times have changed and it is possible that more women have accumulated more assets individually, this situation may still be true for many couples. 245D. De’Armond et al. / Financial Services Review 23 (2014) 239–248 There are unique problems if the surviving spouse remarries. If the executor of the decedent’s estate makes the election to make the unused exclusion available to the surviving spouse’s estate and (s)he remarries, his or her estate will only be able to use the first spouse’s additional exclusion if (s)he dies before the death of the new spouse. There may also be family members concerned that the new spouse may acquire family assets upon the surviving spouse’s death. The solution is for the surviving spouse to make gifts to the children of the first marriage sufficient to use up the unused exclusion of the first decedent’s estate. The gifts must be completed before the death of the new spouse. 4.4. Example Mr. and Mrs. Smith have net assets of $7,500,000, owned equally by each spouse. Mr. Smith dies in 2014, leaving all of his assets directly to Mrs. Smith. Mr. Smith’s estate will have no net estate because of the marital deduction. The executor of Mr. Smith’s estate should make the election to make the unused exclusion available to Mrs. Smith’s estate. If Mrs. Smith remarries, she should make consider making gifts equal to Mr. Smith’s unused exemption. If Mr. Smith died in 2014, the unused exemption would be $5,340,000 and gifts of that amount would leave her with assets of only $2,160,000. The desire to use tax benefits should not overshadow good financial planning. She should not make gifts to the point that that she excessively depletes her assets. 4.5. Should a protective election be made? In a situation where the combined net assets of a couple are less than $5,340,000, the question arises as to the advisability of making the election upon the death of one of the individuals. It is expensive to file a 706–Estate Tax Return. However, it will be expensive if the surviving spouse’s assets increase (somebody wins the lottery every time) and the second spouse ends up with a net estate greater than $5,340,000. A protective election should be carefully considered. A situation where no election is filed and the second-to-die spouse’s estate ends up paying taxes could result in a lawsuit concerning the executor’s fiduciary responsibility to consider the election. 4.6. Example Mr. and Mrs. Smith have net assets of $7,500,000, owned equally by each spouse. Mr. Smith dies in 2014, leaving all of his assets to the couple’s children. Because Mrs. Smith retains only $3,750,000 it appears that her exemption will cover her estate. The executor does not make the election. Mrs. Smith lives on her social security and is able to invest her assets well. She dies with an estate of $6,000,000—more that the exemption amount at that time. The executor should have made the protective election. Reg. Sec. 20.2010–2T(a)(7)(ii) provides some relief for executors in this situation. For items reported on Schedules A – Real Estate, B – Stocks and Bonds, C – Mortgages, Notes and Cash, D – Insurance on the Decedent’s Life, E – Jointly Owned Property, F – Other Miscellaneous Property, G – Transfers During Decedent’s Life, H – Powers of Appointment and I – Annuities the 246 D. De’Armond et al. / Financial Services Review 23 (2014) 239–248 executor is allowed to estimate the value of the assets rather than acquire an acceptable appraisal. All other amounts on the return are subject to the normal rules of filing the 706. This provision significantly reduces the time and effort required to file a normal 706. 4.7. Microsoft Excel model To further explain and provide examples, we have created a Microsoft Excel model, see the Appendix, Portability of Deceased Spouse Unused Exclusion Model. The model serves the purpose of allowing the reader to understand the numbers more specifically by entering information and numbers to exhibit how the basic premise of portability functions relative to other pertinent variables interact and function. The Appendix includes additional notes pertaining to the model as well as a publicly available (live) download link. 5. Conclusion Time is of the essence when detangling tax issues and implications. It is with no doubt that the Act has had lasting effects on how individuals learn, teach, and consider tax policies and provisions as they arise from tax code revision. The underexposed tax policies tend to provide excellent opportunity for academic exploration in and out of the classroom. As Eyssell wrote in 1999, there is credibility and viability in “learning by doing,” such that pedagogy should use the creativity of teaching to further explore these avenues as they occur (Eyssell, 1999). In the case of using the portability election of deceased spouses, relatively little has been written to this point, and even less attention has been turned to classroom curricula encompassing such opportunities, as for the most part, this is new and unchartered territory. Although we are in the first stages of the post-cliff era, many future decisions made at governmental and regulating body levels remain to be made, and, each decision carries the potential to create changes along the way. The portability provision was extended and should be included in estate planning. Accounting, law, and financial planning educators, especially within the tax courses, have a great opportunity to explore the exciting and changing realm of tax code with students given the post-cliff decisions, scenarios, implications, and conclu- sions. Appendix: Portability of deceased spouse unused exclusion model Notes about Excel model: available for public download at http://bit.ly/1tfm97h. The Excel model is designed to illustrate the operation of the portability provisions of the Act in the most basic terms. Readers can input values in the highlighted fields, and the model will make calculations that demonstrate how portability operates on a couple’s estate. The model is designed for illustrative and educational purposes only and does not represent an exhaustive calculation of the estate tax liability for an estate. Of particular note, the model requires certain assumptions for the purpose of making 247D. De’Armond et al. / Financial Services Review 23 (2014) 239–248 calculations. Specifically, the model does not take into account lifetime gifts and their impact on the estate tax, and thus, assumes that no lifetime gifts have been made. Additionally, the model assumes that property is owned by spouses equally and does not give any consider- ation to the impact of state property law regimes like community property. Finally, in the interest in comparing “apples to apples”—more specifically 2024 dollars to 2024 dollars— the model makes an inflationary calculation to the exemption amount available in 2014. The model provides a growth calculation for the surviving spouse’s estate and the surviving spouse’s life expectancy. As a result, the model must inflate the exemption amount to the year of surviving spouse’s predicted death. This inflated amount provides a “best guess” of what the exemption may look like in a future year. In actuality, the IRS acts each year to set the exemption amount based on an inflationary adjustment then rounds to the nearest $1,000. Even with the above assumptions, the model provides an effective tool for illustrating portability and its effect on a couple’s estate. References Beierlein, J. J., & Neverett, M. (2013). Who takes personal finance? Financial Services Review, 22, 151–171. Deener, J. (2012). New portability rules: A cure for incomplete estate planning. Journal of Accountancy, 214, 48–52. Dudley, S. (2013). Fear the regulatory cliff, too. Reason, 44, 31–32. Dunn, D. V. (2011). Bypass the bypass trust? Trusts & Estates, 150, 24. Eyssell, E. (1999) Learning by doing: Offering a university practicum in personal financial planning. Financial Services Review, 8, 293–303. Gallo, J. J. (2011). Why portability isn’t a cure-all. Journal of Financial Planning, 24, 32–33. Grunwald, M. (2012). How to avoid the fiscal cliff. Time, 180, 40. Internal Revenue Service. (2011). Guidance on electing portability of deceased spousal unused exclusion amount. Internal Revenue Bulletin, 2011, 516–518. Internal Revenue Code Sec. 2001(c), as amended by 2012 Taxpayer Relief Act §101(c)(1) Code Sec. 2010(c)(4)(B)(i), as amended by 2012 Taxpayer Relief Act §101(c)(2). Joseph, B. M. (2012). Beware the portability election. California CPA, 80, 13–16. Katz, R. D. (2011). What planners need to know about portability. CPA Journal, 81, 50. Miller, J. A. (2013). Wealth transfer tax planning for 2013 and beyond. Brigham Young University Law Review, 2013, 878. Suderman, P. (2013). The fiscal cliff. Reason, 44, 26–28. 248 D. De’Armond et al. / Financial Services Review 23 (2014) 239–248 Does active management work? Evidence from equity sector funds Crystal Y. Lina,* aSchool of Business, Eastern Illinois University, Charleston, IL 61920, USA Abstract This study presents considerable evidence that equity sector mutual funds, the nine Fidelity Select Portfolios here, have provided better after-expense returns against broader market ETF, SPY, and their peer sector ETFs, the nine Select Sector SPDR Funds, over the sample period 1999–2010. Not only do they achieve higher nominal returns over the 12 years, except for few sector mutual funds, some of the funds also generate higher risk-adjusted returns measured by Sharpe Ratio and ! from various asset pricing models. More important, none of the sector mutual funds generates a significant negative ! for the sample period no matter that asset pricing model is used. The results suggest that actively managed sector funds be considered by individual investors and/or their financial planners for mutual fund selection. © 2014 Academy of Financial Services. All rights reserved. JEL classification: G11; G12 Keywords: Mutual fund performance; Active mutual fund management; Sector investing 1. Introduction The mutual fund industry has seen tremendous growth in the past few decades. The 2013 Investment Company Institute fact book shows that 44.4% of U.S. households owned mutual funds in 2012, up significantly from 4.6% in 1980 and 23.4% in 1990, modestly down from 48.6% in 2000. Bulk of individual’s mutual fund assets, $5.4 trillion out of $11.1 trillion, are invested in equities. Individual accounts hold 90.3% of total equity mutual fund assets ($5.9 trillion). With the expansion of Defined Contribution Plan assets, almost threefold from $1.7 * Corresponding author. Tel.: !1-217-581-2227; fax: !1-217-581-6642. E-mail address: cylin@eiu.edu (C.Y. Lin) Financial Services Review 23 (2014) 249–271 1057-0810/14/$ – see front matter © 2014 Academy of Financial Services. All rights reserved. trillion in 1995 to $5.1 trillion in 2012, knowledge in equity mutual funds becomes more and more important for individual investors. Today many retirement savings plans offer a broader range of investment vehicles for individual investors. Sector funds appear on the investment menu for many plan participants. Do they deserve individual investors’ attention? How is their historical performance against their benchmarks? Should financial planners recommend such mutual funds to their clients? Using sector funds with the longest history, the Fidelity Select Portfolios and the Select Sector SPDR Funds, this article tries to answer these questions for individual investors and/or their financial planners. Once investors decide to allocate their assets to the U.S. public equity market, they must evaluate possible ways to implement the allocation in their subportfolios. There are two dimensions of this implementation. The first one is indexing or not: Do investors want to passively invest so that their returns closely track selected indexes and at the same time investors pay less fees and experience less turnover? Or do investors want to actively manage their subportfolios either in-house or through external fund managers? Sullivan and Xiong (2012) estimate that $1.2 trillion out of $3.5 trillion assets in the U.S. equity mutual funds and ETFs was passively managed as of September 2010; equity index mutual funds and equity ETFs split the share of passively managed equity index funds.1 The second dimension of this implementation is whether to make allocation decisions at the sector level: Do investors want to further divide stocks by sector/industry and set up weight limits to these groups? Or do investors not care about sector issues at all. Porter (1985) and McGahan and Porter (1997) make a strong case for sector investing: They demonstrate that a company’s performance is influenced by the growth and structure of its industry. In addition, Groysberg et al., (2011) show that forecasted industry growth is the most important explanatory variable when analysts construct their forecasts on companies. In this article, I examine the performance of actively managed equity sector funds and their passively managed counterparties to assist investors in their decisions in implementing equity sector asset allocations. The research question I try to answer is: When an investor wants to use external fund managers to allocate assets among U.S. equity sectors, are sector index funds a better choice than actively managed funds or vice versa? Two parallel analyses regarding equity sector fund performance are provided in this article. One is their performance against a broad U.S. equity market index fund, which tests the Efficient Market Hypothesis (EMH). This analysis serves as a general empirical study on U.S. equity market efficiency. The other is sector mutual funds’ performance against sector index funds, which tests EMH in a smaller territory: equity sector. It is arguable that using a broad equity market index is not appropriate when evaluating a sector fund because it has a much smaller universe of securities from which to choose. An index for the same sector would be more appropriate when used as a benchmark. In reality, this is what many actively managed funds do: they select an index that best matches their investment universe as their benchmark. 2. Sector mutual funds and ETFs A variety of equity sector funds have been introduced to the market place during the past several decades. These sector funds allow investors to custom tailor asset allocations to fit 250 C.Y. Lin / Financial Services Review 23 (2014) 249–271 their particular investment needs or goals. Like their broad equity market counterparties, index sector funds emerged later than actively managed sector funds. Two groups of sector funds with the longest history in each category are used in this study: the Fidelity Select Portfolios and the Select Sector SPDR Funds. In addition, SPDR S&P 500 ETF is used as an investable benchmark for the study. 2.1. Fidelity select portfolios The Fidelity’s Web site2 listed 38 mutual funds under its Stock Funds/Sector Funds category at the time of writing this article. The inception date of the earliest three sector funds (Energy, Healthcare, and Technology) is July 14, 1981. I identified nine broader sector funds based on fund prospectus: Select Consumer Discretionary Portfolio (FSCPX), Select Con- sumer Staples Portfolio (FDFAX), Select Energy Portfolio (FSENX), Select Financial Services Portfolio (FIDSX), Select Health Care Portfolio (FSPHX), Select Industrials Port- folio (FCYIX), Select Materials Portfolio (FSDPX), Select Technology Portfolio (FSPTX), and Select Utilities Portfolio (FSUTX).3 The above funds match nine sector index funds that are described later. These sector mutual funds were started between July 1981 and March 1997. Most of the other funds listed are narrower focused industry funds. According to Fidelity’s fund prospectuses, these funds seek capital appreciation, invest in domestic and foreign issuers, normally invest primarily in common stocks, and invest at least 80% of assets in securities of companies principally engaged in the selected sector. In other words, these funds are actively managed. Fidelity Management & Research Company is the fund’s manager. The Fidelity Select Portfolios have an expense ratio between 0.80% (Healthcare) and 1.41% (Materials) as of February 29, 2012. The portfolio turnover rate is between 35% (Consumer Staples) and 384% (Financial Services). The funds have net assets between $0.28 billion (Consumer Discretionary) and $2.50 billion (Energy).4 Fidelity charges a Short-Term Redemption Fee, 0.75%, when money is withdrawn from a sector fund within 30 days of purchase to reduce short-term mutual fund trading.5 All these sector funds are open to new investors. 2.2. Select Sector SPDR Funds The Select Sector SPDR Trust was organized as a Massachusetts business trust on June 10, 1998. State Street Global Advisors serves as the fund manager. The Trust consists of nine separate investment portfolios (each a “Select Sector SPDR Fund”) incepted in December 1998: The Consumer Discretionary Select Sector SPDR Fund (XLY), The Consumer Staples Select Sector SPDR Fund (XLP), The Energy Select Sector SPDR Fund (XLE), The Financial Select Sector SPDR Fund (XLF), The Health Care Select Sector SPDR Fund (XLV), The Industrial Select Sector SPDR Fund (XLI), The Materials Select Sector SPDR Fund (XLB), The Technology Select Sector SPDR Fund (XLK), and The Utilities Select Sector SPDR Fund (XLU). These sector funds seek to provide investment results that, before expenses, correspond generally to the price and yield performance of publicly traded equity securities of companies in certain “Select Sector Indexes”: The Consumer Discretionary 251C.Y. Lin / Financial Services Review 23 (2014) 249–271 Select Sector Index, The Consumer Staples Select Sector Index, The Energy Select Sector Index, The Financial Select Sector Index, The Health Care Select Sector Index, The Industrial Select Sector Index, The Materials Select Sector Index, The Technology Select Sector Index, and The Utilities Select Sector Index. Each stock in the S&P 500 is allocated to one and only one Select Sector Index. The combined companies of the nine Select Sector Indexes represent all of the companies in the S&P 500. That is, the Select Sector SPDR Funds unbundle the S&P 500.6 These passively managed sector funds use a replication strategy, attempting to track the performance of an unmanaged index of securities. According to Select Sector SPDR Fund Annual Report, the ratio of expenses to average net assets is 0.19% for each individual fund as of September 30, 2011. The turnover rate is between 3.20% (Utilities) and 13.86% (Materials). The funds have net assets between $1.64 billion (Materials) and $6.64 billion (Utilities).7 2.3. SPDR S&P 500 ETF The SPDR S&P 500 ETF (SPY) is an exchange traded fund designed to generally correspond to the price and yield performance of the S&P 500 Index. Utilizing a full replication approach, the Trust owns all 500 securities of the S&P 500 Index in their approximate market capitalization weight. The fund was incepted on January 22, 1993. The portfolio has an expense ratio of 0.09%, a turnover rate of 3.72%, and $80.87 billion net assets as of September 30, 2011.8 Both equity mutual funds and ETFs are pooled investments that represent ownership in a basket of stocks. However, ETFs can be traded like individual stocks. They are also shortable, marginable, and optionable. Index ETFs normally have lower fees by eliminating many of the operating, research, and transaction expenses incurred by active money man- agers. They also provide greater transparency: one can get a holding list more frequently than with mutual funds. For example, Fidelity Select Portfolios publish monthly holdings whereas the Select Sector SPDRs update their online information daily. 3. A first look at equity sector funds: raw returns My analysis starts in January 1999 and ends in December 2010,9 since the earliest price data available for Select Sector SPDR Funds is mid-December 1998. I downloaded price and dividend data from Yahoo!Finance Web site.10 Monthly, annual, and 12-year holding period return is calculated for each fund as: Ri,t " Pi,t # Di,t Pi,t"1 $ 1, (1) where Ri,t is the return for fund i during period t, Pi,t is the price for fund i at the end of period t, Pi,t"1 is the price for fund i at the end of period t " 1, and Di,t is the total dividend/cash distribution of fund i during period t. 252 C.Y. Lin / Financial Services Review 23 (2014) 249–271 The purpose of this study is to assist investors implement asset allocations at the sector level, therefore, I use fund price instead of fund Net Asset Value to calculate fund return. This return is net of expenses and is attainable. As argued by Jones and Wermers (2011), I compare actively managed sector mutual fund performance to their passive alternative and not to the index itself. For the same reason, I use SPY as an investable broad U.S. equity market benchmark. 3.1. Twelve-year return Which group of funds generates higher returns during the sample period? The results are summarized in Table 1. In the rest of the article, I use MF to represent Fidelity Select Portfolios, ETF to represent Select Sector SPDR Funds, and SPY to represent the SPDR S&P 500 ETF for easier reference. Panel A of Table 1 lists the 12-year (1999–2010) holding period return for the 19 funds. The highest return is 295.2% for the Energy MF and the lowest return is "17.4% for the Technology ETF, and the return for SPY is 20.8%. Seven of nine MFs, except for the Utilities and Consumer Discretionary MF, have a higher return than that of their ETF counterparties; the average outperformance is 51.2%. When SPY is used as the benchmark, eight out of nine MFs (except Utilities) and seven out of nine ETFs (except Financial and Technology) outperform. The average MF holding period return across all sectors is 109.2% versus 58.0% for ETFs, and the difference is significant at the 5% level using one-tailed t test. Results are slightly different when I compound annual holding period returns through the 12 years. Panel B of Table 1 shows that the only underperforming MF against SPY is Financial, not Utilities. The same seven MFs beat peer ETFs with an average outperformance of 63.3%. The average MF 12-year return with annual compounding across all sectors is 127.6% versus 64.3% for ETFs, and the difference is also significant at the 5% level using one-tailed t test. These results show that on average sector MFs outperform their ETF counterparties for the whole sample period. None of the nine MFs suffers a loss during the sample period; however, the Financial and Technology ETFs generate a negative return under both calculation methods. The tech bubble in early 2000s and the financial crisis in 2008 contribute to the negative 12-year return for these two sectors. Although sector ETFs mimicked their benchmark indexes during these bear markets, it seems sector MF managers made the right decisions against trend changes. Fig. 1 shows that the annual return for the Technology MF is 119.07%, "28.18%, and "31.70% for the year 1999, 2000, and 2001, whereas the annual return for its peer ETF is 65.13%, "41.89%, and "23.34%, respectively. The annual return for the Financial MF is "13.42%, "49.83%, and 23.71% for the year 2007, 2008, and 2009, whereas the annual return for its peer ETF is "18.89%, "54.06%, and 16.98%, respectively. 3.2. Average annual return Seven out of nine MFs (except for Utilities and Consumer Discretionary) outperform their peer ETFs when measured with average annual holding period return depicted in Fig. 2. An interesting finding in Fig. 2 is that the ranking of performance is different for MFs and ETFs. Energy MF has the highest average annual return of 18.3%, followed by Materials (17.0%), 253C.Y. Lin / Financial Services Review 23 (2014) 249–271 Technology (14.9%), Industrials (11.5%), Consumer Staples (7.7%), Healthcare (6.3%), Utilities (4.7%), Consumer Discretionary (4.6%), and Financial (3.4%). In the ETF group, Energy (13.8%) and Materials (9.7%) are still ranked first and second, whereas Financial (1.9%) is again at the bottom. However, the other six sectors are ranked differently. Fig. 2 also shows that for the Materials, Industrials, Technology, and Consumer Staples sector, MF average annual holding period returns are higher than that of corresponding ETFs at the 1% or 5% significance level using one-tailed t test. Table 1 Twelve-year return, 1999–2010 Sector MF ETF MF beats ETF Difference MF beats SPY ETF beats SPY A: Holding period return B 287.6% 105.1% Yes 182.5% Yes Yes E 295.2% 221.3% Yes 73.9% Yes Yes F 23.4% "8.0% Yes 31.4% Yes No I 125.9% 62.1% Yes 63.8% Yes Yes K 73.5% "17.4% Yes 90.9% Yes No P 82.0% 26.4% Yes 55.7% Yes Yes U 20.6% 41.3% No "20.7% No Yes V 50.4% 36.2% Yes 14.2% Yes Yes Y 24.0% 55.0% No "30.9% Yes Yes Average fund return 109.2% 58.0% p-value average MF return higher than average ETF return 0.022 SPY 20.8% B: Return with annual compounding B 325.9% 118.7% Yes 207.2% Yes Yes E 354.3% 243.2% Yes 111.0% Yes Yes F 8.4% "13.7% Yes 22.0% No No I 148.8% 67.8% Yes 81.0% Yes Yes K 70.5% "16.2% Yes 86.7% Yes No P 111.2% 31.6% Yes 79.6% Yes Yes U 25.7% 51.8% No "26.1% Yes Yes V 67.6% 37.4% Yes 30.1% Yes Yes Y 36.4% 58.4% No "21.9% Yes Yes Average fund return 127.6% 64.3% p-value average MF return higher than average ETF return 0.015 SPY 23.9% Note: B represents the Materials sector; E represents the Energy sector; F represents the Financial sector; I represents the Industrials sector; K represents the Technology sector; P represents the Consumer Staples sector; U represents the Utilities sector; V represents the Healthcare sector; Y represents the Consumer Discretionary sector; and SPY represents the SPDR S&P 500 Trust. Holding period return in Panel A is calculated by adding ending price and all dividends paid in 12 years then divided by beginning price. Return with annual compounding in Panel B is calculated by compounding annual holding period returns for each individual fund. 254 C.Y. Lin / Financial Services Review 23 (2014) 249–271 -100% -50% 0% 50% 100% Materials -60% -40% -20% 0% 20% 40% 60% Energy -60% -40% -20% 0% 20% 40% 60% Industrials -40% -20% 0% 20% 40% Consumer Staples -60% -40% -20% 0% 20% 40% Healthcare -60% -40% -20% 0% 20% 40% Financial -60% -40% -20% 0% 20% 40% 60% 80% 100% 120% 140% Technology -40% -20% 0% 20% 40% Utilities -60% -40% -20% 0% 20% 40% 60% Consumer Discretionary MF ETF SPY Fig. 1. Annual holding period return. 255C.Y. Lin / Financial Services Review 23 (2014) 249–271 When annual holding period return is compared between MFs and ETFs, for the same sector, Table 2 shows that seven out of nine MFs (except Utilities and Consumer Discre- tionary) generate a higher return in seven or more years in the 12 year sample period. The average number of years of outperforming is 7.9 and the percentage of years of outperform- ing is 66%. Using SPY as the benchmark, on average in 7.4 out of 12 years MFs beat SPY (62% of years); only in 6.4 years does the ETF group beat SPY (54% of years). Table 2 Annual return comparison, 1999–2010 Sector MF beats ETF MF beats SPY ETF beats SPY Number of years % of years Number of years % of years Number of years % of years B 9 75% 10 83% 9 75% E 7 58% 9 75% 7 58% F 9 75% 6 50% 7 58% I 10 83% 9 75% 7 58% K 8 67% 6 50% 4 33% P 10 83% 9 75% 5 42% U 6 50% 6 50% 6 50% V 7 58% 7 58% 5 42% Y 5 42% 5 42% 8 67% Average 7.9 66% 7.4 62% 6.4 54% Note: B represents the Materials sector; E represents the Energy sector; F represents the Financial sector; I represents the Industrials sector; K represents the Technology sector; P represents the Consumer Staples sector; U represents the Utilities sector; V represents the Healthcare sector; Y represents the Consumer Discretionary sector; and SPY represents the SPDR S&P 500 Trust. Total number of years: 12. P-value Average MF annual holding period return higher than average ETF annual holding period return B E F I K P U V Y 0.012 0.082 0.092 0.002 0.046 0.013 0.430 0.300 0.222 0% 5% 10% 15% 20% B E F I K P U V Y MF ETF SPY Fig. 2. Arithmetic mean of annual holding period return. B represents the Materials sector; E represents the Energy sector; F represents the Financial sector; I represents the Industrials sector; K represents the Technology sector; P represents the Consumer Staples sector; U represents the Utilities sector; V represents the Healthcare sector; Y represents the Consumer Discretionary sector; and SPY represents the SPDR S&P 500 Trust. 256 C.Y. Lin / Financial Services Review 23 (2014) 249–271 3.3. Decomposition of 12-year holding period return What portion of that 12-year holding period return is contributed by capital gains? What portion is contributed by dividend yield (regular dividend and special cash distribution)? There are four MFs (Financial, Utilities, Healthcare, and Consumer Discretionary) and two ETFs (Financial and Technology) have negative capital gains during the period. The SPY has a capital gains yield of 2.0%. Actually, only three MFs (Materials, Energy, and Industrials) have a higher capital gains yield than dividend yield. That number is five for ETFs (Materials, Energy, Industrials, Healthcare, and Consumer Discretionary). The SPY has a dividend yield of 18.8%. With 2.0% capital gains yield from SPY, the 12-year sample period is pretty flat. The S&P 500 price index has two peaks, 1552.87 on March 24, 2000 and 1576.09 on October 17, 2007, and two troughs, 768.63 on October 10, 2002 and 666.79 on March 6, 2009. This range provides a good testing field for performance analysis. After examining 12-year returns and average annual returns, I find that most sector MFs outperform both their ETF peers and SPY for most years during the 1999–2010 sample period. The exceptions are the Utilities and Consumer Discretionary MF. 4. A closer look at equity sector funds: risk adjusted returns Before drawing a conclusion on sector MFs’ performance, one must investigate the risk dimension of the returns. Here, I look at risk adjusted returns that incorporate both total risk and systematic risk based on monthly holding period returns. Table 3 summarizes statistics of both monthly holding period returns and excess returns, which are calculated as monthly holding period return minus monthly 1-month T-bill rate. Panel A shows the maximum monthly holding period return is 31.446% (Technology MF, February 2000) and the minimum monthly return is "28.379% (Technology MF, February 2001). Because sector funds focus on specific investment areas, it is expected that both sector MFs and ETFs have a higher volatility when compared to a broader market benchmark. That Table 3 Summary statistics of monthly return (%) Maximum Minimum Mean Median SD Skewness Kurtosis N Sharpe Ratio A: Raw return MF 31.446 "28.379 0.695 1.014 6.316 "0.238 2.738 1289 ETF 24.768 "26.198 0.468 0.796 5.982 "0.255 1.762 1296 SPY 9.935 "16.519 0.260 0.741 4.652 "0.485 0.594 144 Overall 31.446 "28.379 0.564 0.897 6.081 "0.246 2.326 2729 B: Excess return MF 31.016 "28.769 0.474 0.759 6.321 "0.229 2.696 1289 0.075 ETF 24.628 "26.198 0.245 0.533 5.985 "0.246 1.730 1296 0.041 SPY 9.925 "16.599 0.038 0.526 4.660 "0.454 0.535 144 0.008 Overall 31.016 "28.769 0.342 0.652 6.085 "0.236 2.288 2729 0.056 257C.Y. Lin / Financial Services Review 23 (2014) 249–271 is true as shown in Table 3. The sector MFs have the widest span of monthly returns (59.825%), followed by sector ETFs (50.966%); both are much higher than that of SPY (26.454%). The standard deviation for sector MF, ETF, and SPY is 6.316%, 5.982%, and 4.652%, respectively. All fund returns are negatively skewed. These characteristics are similar in excess returns as presented in Panel B. The group Sharpe Ratio is 0.075, 0.041, 0.008, and 0.056 for the MFs, ETFs, SPY, and overall funds, respectively. The sector MF group has the highest Sharpe Ratio. These small but positive Sharpe Ratios reflect the flat U.S. equity market during the sample period. 4.1. Sharpe Ratio Does the Sharpe Ratio comparison between each pair of sector funds echo the group Sharpe Ratio results in Table 3? Table 4 shows that for the full sample period, January 1999 through December 2010, seven out of nine sector MFs have a higher Sharpe Ratio than that of their peer ETFs, except for the Utilities and Consumer Discretionary MF. The Materials Table 4 Individual fund Sharpe ratio, January 1999 through December 2010 Fund Mean excess return SD Sharpe Ratio MF SR higher than ETF MF SR higher than SPY ETF SR higher than SPY BBB 1.030 6.786 0.152 Yes Yes XLB 0.555 6.760 0.082 Yes EEE 1.100 7.390 0.149 Yes Yes XLE 0.853 6.504 0.131 Yes FFF 0.035 6.224 0.006 Yes No XLF "0.089 6.886 "0.013 No III 0.669 6.157 0.109 Yes Yes XLI 0.314 5.868 0.054 Yes KKK 0.671 10.126 0.066 Yes Yes XLK "0.015 8.069 "0.002 No PPP 0.367 3.728 0.098 Yes Yes XLP 0.042 3.709 0.011 Yes UUU 0.066 5.021 0.013 No Yes XLU 0.184 4.671 0.039 Yes VVV 0.224 4.175 0.054 Yes Yes XLV 0.093 4.280 0.022 Yes YYY 0.112 4.896 0.023 No Yes XLY 0.272 5.899 0.046 Yes SPY 0.038 4.660 0.008 Average MF SR 0.074 Average ETF SR 0.041 p-value average MF SR higher than average ETF SR 0.020 Note: BBB, EEE, FFF, III, KKK, PPP, UUU, VVV, and YYY represent the Fidelity Select Portfolio mutual funds for the Materials sector, the Energy sector, the Financial sector, the Industrials sector, the Technology sector, the Consumer Staples sector, the Utilities sector, the Healthcare sector, and the Consumer Discretionary sector, respectively. XLB, XLE, XLF, XLI, XLK, XLP, XLU, XLV, and XLY represent the Select Sector SPDR ETFs for these nine sectors, respectively. SPY represents the SPDR S&P 500 Trust. 258 C.Y. Lin / Financial Services Review 23 (2014) 249–271 and Energy MF have the highest Sharpe Ratios, 0.152 and 0.149, respectively; whereas the Financial and Technology ETF have a negative Sharpe Ratio for the same period. The average MF Sharpe Ratio across all sectors is 0.074 versus 0.041 for sector ETFs, and the difference is significant at the 5% level using one-tailed t test. When the sample period is divided into two equal-length subperiods, the results are slightly different.11 For the first half sample period, from January1999 through December 2004, the Healthcare MF does not have a higher Sharpe Ratio than its peer ETF. For the second half sample period, from January 2005 through December 2010, the Energy MF does not have a higher Sharpe Ratio, but the Consumer Discretionary MF does outperform its peer ETF. Subsample analysis also shows that the significant higher MF average Sharpe Ratio for the whole sample period is mainly because of higher MF average Sharpe Ratio in the second half sample period, which is significant at the 1% level, whereas the p-value is greater than 5% for the first half sample period. Table 4 also compares Sharpe Ratio between individual sector funds and SPY. Only the Financial MF underperforms SPY for the full sample period, whereas both the Financial and Technology ETF underperform SPY for the same time period. Using total risk as the measurement, I find that most sector MFs outperform their peer ETFs. Sector MFs also have a higher number of funds outperform SPY. 4.2. Performance against S&P 500 ETF Systematic risk is always the part of risk that gets more attention because many argue that unsystematic risk can be diversified away at a relatively low cost.12 CAPM has been the standard model to test fund performance. I modify the model by replacing excess market return with excess SPY return because an investable benchmark makes more sense for comparing attainable returns: Ri,t $ Rf,t " !i # %i#RSPY,t $ Rf,t$ # &i,t (2) where Ri,t is the return of fund i in month t, Rf,t is the return of one-month T-bill in month t, RSPY,t is the SPY return in month t, and &i,t is an error term. Table 5 shows that the Materials MF generates a 0.988% monthly abnormal return (11.856% annually) and the Energy MF generates a 1.064% monthly abnormal return (12.768% annually) during the sample period, significant at the 1% and 5% level, respec- tively. The Industrial ETF has a significant ! of 0.672% at the 1% level. All %s are significant at the 1% level. Both MFs and ETFs for the Materials sector, the Financial sector, the Industrials sector, and the Technology sector have a % greater than 1. All other funds have a % less than 1 except for the Consumer Discretionary ETF. The average Adjusted R2 is 0.563 for the MFs and 0.555 for the ETFs. For the first half sample period, none of the !s is significant at the 5% level. The average Adjusted R2 declines to 0.438 for the MFs and 0.461 for the ETFs. The average Adjusted R2 are higher for the second half sample period: 0.719 for the MFs and 0.692 for the ETFs. The Materials MF, the Consumer Staples MF, and the Industrials ETF generate 0.940%, 0.495%, and 0.543% monthly abnormal returns, respectively, from January 2005 through December 2010, at the 5% level. 259C.Y. Lin / Financial Services Review 23 (2014) 249–271 Fama-French three-factor model is also modified by replacing excess market return with excess SPY return: Ri,t $ Rf,t " !i # %i#RSPY,t $ Rf,t$ # hiHMLt # siSMBt # &i,t (3) where HML (High minus Low) is the average return on two value portfolios minus the average return on two growth portfolios and SMB (Small minus Big) is the average return on three small portfolios minus the average return on three big portfolios (Fama and French 1993). The data are downloaded from Kenneth R. French Data Library.13 Results in Table 6 show that over the full sample period two sector MFs (Materials and Technology) have a positive ! at the 5% level. None of the ETFs has a significant ! at the 5% level. All %s are significant at the 1% level with the same above/below 1 % distribution as in Table 5. HML is not significant for the Healthcare MF and ETF, and it is significantly negative for three funds (Technology MF and ETF, and Utilities MF). The SMB, size factor, is the least significant factor of the three: only seven out of 18 funds have a significant coefficient. During the first half sample period, none of the MFs or ETFs generates a positive ! at the 5% level. Fewer HML and SMB coefficients are significant. During the second half sample Table 5 One factor model results using SPY excess return, January 1999 through December 2010 Fund ! % Adjusted R2 Estimate SE t-value Estimate SE t-value BBB 0.988 0.362 2.727*** 1.121 0.078 14.379*** 0.590 XLB 0.512 0.347 1.475 1.145 0.075 15.321*** 0.620 EEE 1.064 0.491 2.164** 0.962 0.106 9.086*** 0.363 XLE 0.821 0.436 1.882 0.834 0.094 8.880*** 0.352 FFF "0.005 0.309 "0.018 1.074 0.067 16.126*** 0.644 XLF "0.134 0.345 "0.387 1.183 0.074 15.921*** 0.638 III 0.272 0.230 1.183 1.164 0.052 22.589*** 0.789 XLI 0.672 0.242 2.782*** 1.112 0.050 22.444*** 0.778 KKK 0.606 0.522 1.162 1.711 0.112 15.228*** 0.617 XLK "0.071 0.359 "0.197 1.466 0.077 18.977*** 0.715 PPP 0.349 0.248 1.407 0.485 0.053 9.092*** 0.363 XLP 0.025 0.256 0.097 0.451 0.055 8.186*** 0.315 UUU 0.035 0.278 0.127 0.807 0.060 13.463*** 0.557 XLU 0.165 0.339 0.486 0.497 0.073 6.807*** 0.240 VVV 0.202 0.268 0.755 0.574 0.058 9.950*** 0.406 XLV 0.066 0.223 0.296 0.718 0.048 14.963*** 0.609 YYY 0.078 0.209 0.374 0.903 0.045 20.058*** 0.737 XLY 0.231 0.256 0.902 1.082 0.055 19.611*** 0.728 Note: BBB, EEE, FFF, III, KKK, PPP, UUU, VVV, and YYY represent the Fidelity Select Portfolio mutual funds for the Materials sector, the Energy sector, the Financial sector, the Industrials sector, the Technology sector, the Consumer Staples sector, the Utilities sector, the Healthcare sector, and the Consumer Discretionary sector, respectively. XLB, XLE, XLF, XLI, XLK, XLP, XLU, XLV, and XLY represent the Select Sector SPDR ETFs for these nine sectors, respectively. SPY represents the SPDR S&P 500 Trust. ***Significant at the 1% level. **Significant at the 5% level. 260 C.Y. Lin / Financial Services Review 23 (2014) 249–271 T ab le 6 T hr ee fa ct or re gr es si on re su lts us in g SP Y ex ce ss re tu rn ,J an ua ry 19 99 th ro ug h D ec em be r 20 10 Fu nd ! % H M L SM B A dj us te d R 2 E st im at e SE t- va lu e E st im at e SE t- va lu e E st im at e SE t- va lu e E st im at e SE t- va lu e B B B 0. 71 9 0. 34 0 2. 11 7* * 1. 14 5 0. 07 2 15 .9 01 ** * 0. 50 4 0. 09 3 5. 39 6* ** 0. 13 7 0. 09 1 1. 50 0 0. 65 6 X L B 0. 32 7 0. 33 0 0. 99 0 1. 17 4 0. 07 0 16 .7 50 ** * 0. 43 0 0. 09 1 4. 73 6* ** 0. 04 0 0. 08 9 0. 44 8 0. 67 2 E E E 0. 92 5 0. 49 7 1. 86 1 0. 98 2 0. 10 5 9. 31 4* ** 0. 31 4 0. 13 7 2. 29 9* * 0. 03 6 0. 13 4 0. 26 6 0. 37 9 X L E 0. 73 3 0. 43 8 1. 67 5 0. 86 1 0. 09 3 9. 27 3* ** 0. 29 6 0. 12 0 2. 45 7* * " 0. 04 1 0. 11 8 " 0. 35 1 0. 37 8 FF F " 0. 20 9 0. 24 7 " 0. 84 5 1. 12 4 0. 05 2 21 .4 22 ** * 0. 59 5 0. 06 8 8. 74 7* ** " 0. 03 6 0. 06 6 " 0. 54 0 0. 78 3 X L F " 0. 35 2 0. 26 5 " 1. 32 6 1. 24 3 0. 05 6 22 .0 85 ** * 0. 68 7 0. 07 3 9. 41 2* ** " 0. 07 1 0. 07 1 " 0. 99 0 0. 79 6 II I 0. 37 3 0. 20 5 1. 81 8 1. 17 2 0. 04 3 27 .3 37 ** * 0. 45 4 0. 05 6 8. 05 7* ** 0. 17 7 0. 05 5 3. 20 4* ** 0. 85 6 X L I 0. 11 8 0. 21 2 0. 55 5 1. 13 3 0. 04 5 25 .2 00 ** * 0. 33 8 0. 05 8 5. 79 3* ** 0. 04 7 0. 05 7 0. 83 0 0. 82 0 K K K 0. 61 4 0. 27 1 2. 26 2* * 1. 56 5 0. 05 8 27 .1 83 ** * " 1. 01 4 0. 07 5 " 13 .5 80 ** * 0. 65 0 0. 07 3 8. 90 3* ** 0. 90 1 X L K 0. 12 6 0. 22 4 0. 56 4 1. 38 6 0. 04 7 29 .1 97 ** * " 0. 80 4 0. 06 2 " 13 .0 57 ** * 0. 18 4 0. 06 0 3. 06 1* ** 0. 89 4 PP P 0. 23 9 0. 21 5 1. 11 3 0. 52 2 0. 04 6 11 .4 61 ** * 0. 39 0 0. 05 9 6. 60 9* ** " 0. 06 5 0. 05 8 " 1. 12 4 0. 54 4 X L P 0. 05 1 0. 23 1 0. 21 9 0. 49 2 0. 04 9 10 .0 34 ** * 0. 25 0 0. 06 4 3. 93 6* ** " 0. 20 8 0. 06 2 " 3. 35 5* ** 0. 46 7 U U U 0. 17 5 0. 27 9 0. 62 7 0. 80 0 0. 05 9 13 .5 28 ** * " 0. 22 4 0. 07 7 " 2. 92 4* ** " 0. 09 5 0. 07 5 " 1. 27 0 0. 57 7 X L U 0. 13 9 0. 31 4 0. 44 2 0. 54 5 0. 06 7 8. 18 8* ** 0. 36 0 0. 08 6 4. 17 3* ** " 0. 19 0 0. 08 4 " 2. 25 7* * 0. 38 1 V V V 0. 17 4 0. 27 6 0. 63 0 0. 57 9 0. 05 8 9. 90 5* ** 0. 06 9 0. 07 6 0. 90 9 0. 00 5 0. 07 4 0. 06 5 0. 40 2 X L V 0. 15 6 0. 22 7 0. 68 8 0. 72 6 0. 04 8 15 .0 86 ** * " 0. 05 9 0. 06 2 " 0. 94 7 " 0. 11 8 0. 06 1 " 1. 93 1 0. 61 4 Y Y Y " 0. 18 3 0. 18 7 " 0. 98 0 0. 90 2 0. 04 0 22 .7 62 ** * 0. 32 2 0. 05 1 6. 26 8* ** 0. 24 3 0. 05 0 4. 83 2* ** 0. 80 0 X L Y 0. 03 1 0. 24 6 0. 12 5 1. 09 0 0. 05 2 20 .8 55 ** * 0. 30 9 0. 06 8 4. 55 3* ** 0. 14 6 0. 06 6 2. 19 9* * 0. 76 0 N ot e: B B B ,E E E ,F FF ,I II ,K K K ,P PP ,U U U ,V V V ,a nd Y Y Y re pr es en t th e Fi de lit y Se le ct Po rt fo lio m ut ua l fu nd s fo r th e M at er ia ls se ct or ,t he E ne rg y se ct or , th e Fi na nc ia l se ct or , th e In du st ri al s se ct or , th e T ec hn ol og y se ct or , th e C on su m er St ap le s se ct or , th e U til iti es se ct or , th e H ea lth ca re se ct or , an d th e C on su m er D is cr et io na ry se ct or ,r es pe ct iv el y. X L B ,X L E ,X L F, X L I, X L K ,X L P, X L U ,X L V ,a nd X L Y re pr es en t th e Se le ct Se ct or SP D R E T Fs fo r th es e ni ne se ct or s, re sp ec tiv el y. SP Y re pr es en ts th e SP D R S& P 50 0 T ru st . ** *S ig ni fic an t at th e 1% le ve l. ** Si gn ifi ca nt at th e 5% le ve l. 261C.Y. Lin / Financial Services Review 23 (2014) 249–271 period, three sector MFs (Materials, Industrials, and Consumer Staples) have a significant positive !. The Financial ETF generates a significant negative ! at the 5% significance level. The four-factor model is modified by replacing excess market return with excess SPY return: Ri,t $ Rf,t " !i # %i#RSPY,t $ Rf,t$ # hiHMLt # siSMBt # miMOMt # &i,t (4) where MOM (momentum) is the average return on the two high prior return portfolios minus the average return on the two low prior return portfolios (see Carhart 1997). The data are downloaded from Kenneth R. French Data Library. The results presented in Table 7 resemble those from the three-factor model. Table 7 shows that over the full sample period the same two sector MFs (Materials and Technology) have a positive ! at the 5% level. For the momentum factor, MOM, only five out of 18 funds have a significant coefficient. Generally speaking, more often, sector MFs generate significant higher !s than peer ETFs no matter whether a one-factor, three-factor, or four-factor model is adopted when SPY is used as the proxy for market portfolio. All fund returns are sensitive to the overall equity market movements as the % coefficients for excess SPY return are all significant at the 1% level. Fund returns are less sensitive to the value/growth factor, the size factor, and the momentum factor in the order of listing. The findings on performance against SPY are not consistent with EMH. They support the argument by Kacperczyk, Sialm, and Zheng (2005) that more sector/industry concentrated funds perform better. 4.3. Sector mutual fund performance against sector index funds Let us examine MF performance against peer ETF within each sector. Excess return of peer ETF is used as the independent variable for the one-factor model: RMFj,t $ Rf,t " !j # %j#RETFj,t $ Rf,t$ # &j,t (5) where RMFj,t is the return of MF of sector j in month t, and RETFj,t is the return of ETF in the same sector j in month t. Table 8 reports that three sector MFs (Materials, Industrials, and Technology) generate significant positive !s, 0.506%, 0.446%, and 0.689%, respectively, during the full sample period. That is equivalent to annualized outperformance of 6.072%, 5.352%, and 8.268%, respectively. Subsample analysis shows that none of the sector MFs outperforms in the first half sample period, whereas the Materials and the Industrials MF outperform in the second half sample period. All %s are positive and significant at the 1% level. Which model can explain most of the return variances of sector funds? I summarize the Adjusted R2 for the four models in Table 9. Most of the time, adding HML and SMB does increase the Adjusted R2 when using SPY as the benchmark. Only one out of 18 regressions for the full sample period suffers a slight decrease of explaining power, 0.004. Adding MOM, however, does not increase Adjusted R2 across the board. Over the full sample period, the average Adjusted R2 across all funds is 0.559, 0.649, 0.653, and 0.764 for the one/three/four-factor model using SPY and one-factor 262 C.Y. Lin / Financial Services Review 23 (2014) 249–271 T ab le 7 Fo ur fa ct or re gr es si on re su lts us in g ex ce ss SP Y re tu rn ,J an ua ry 19 99 th ro ug h D ec em be r 20 10 Fu nd ! % H M L SM B M O M A dj us te d R 2 E st im at e SE t- va lu e E st im at e SE t- va lu e E st im at e SE t- va lu e E st im at e SE t- va lu e E st im at e SE t- va lu e B B B 0. 70 6 0. 34 1 2. 07 2* * 1. 17 1 0. 08 1 14 .5 04 ** * 0. 51 5 0. 09 5 5. 43 1* ** 0. 12 8 0. 09 2 1. 38 7 0. 04 2 0. 05 8 0. 71 7 0. 65 5 X L B 0. 31 6 0. 33 2 0. 95 3 1. 19 6 0. 07 9 15 .2 14 ** * 0. 44 0 0. 09 2 4. 76 4* ** 0. 03 2 0. 09 0 0. 35 9 0. 03 5 0. 05 7 0. 62 7 0. 67 E E E 0. 87 8 0. 49 4 1. 77 8 1. 07 7 0. 11 7 9. 19 8* ** 0. 35 6 0. 13 8 2. 58 6* * 0. 00 3 0. 13 4 0. 02 4 0. 15 2 0. 08 4 1. 80 4 0. 38 9 X L E 0. 68 4 0. 43 3 1. 58 2 0. 96 1 0. 10 3 9. 36 2* ** 0. 33 9 0. 12 1 2. 81 4* ** " 0. 07 5 0. 11 7 " 0. 64 3 0. 15 9 0. 07 4 2. 15 9* * 0. 39 4 FF F " 0. 18 8 0. 24 6 " 0. 76 3 1. 08 0 0. 05 8 18 .5 13 ** * 0. 57 6 0. 06 9 8. 40 2* ** " 0. 02 1 0. 06 7 " 0. 31 5 " 0. 07 0 0. 04 2 " 1. 65 8 0. 78 6 X L F " 0. 31 9 0. 26 1 " 1. 22 1 1. 17 7 0. 06 2 18 .9 88 ** * 0. 65 8 0. 07 3 9. 03 7* ** " 0. 04 8 0. 07 1 " 0. 67 7 " 0. 10 6 0. 04 5 " 2. 38 1* * 0. 80 2 II I 0. 37 0 0. 20 6 1. 79 6 1. 17 8 0. 04 8 24 .3 18 ** * 0. 45 6 0. 05 7 7. 99 0* ** 0. 17 4 0. 05 6 3. 11 6* ** 0. 00 9 0. 03 5 0. 25 3 0. 85 5 X L I 0. 13 2 0. 21 2 0. 62 3 1. 10 4 0. 05 0 21 .9 92 ** * 0. 32 5 0. 05 9 5. 51 3* ** 0. 05 7 0. 05 7 0. 99 6 " 0. 04 6 0. 03 6 " 1. 27 3 0. 82 1 K K K 0. 61 0 0. 27 3 2. 23 5* * 1. 57 5 0. 06 5 24 .3 50 ** * " 1. 01 0 0. 07 6 " 13 .2 95 ** * 0. 64 7 0. 07 4 8. 75 1* ** 0. 01 5 0. 04 7 0. 31 6 0. 90 0 X L K 0. 15 9 0. 21 8 0. 72 8 1. 31 9 0. 05 2 25 .4 62 ** * " 0. 83 3 0. 06 1 " 13 .6 80 ** * 0. 20 7 0. 05 9 3. 49 6* ** " 0. 10 7 0. 03 7 " 2. 86 8* ** 0. 89 9 PP P 0. 22 2 0. 21 4 1. 03 9 0. 55 5 0. 05 1 10 .9 42 ** * 0. 40 5 0. 06 0 6. 78 8* ** " 0. 07 6 0. 05 8 " 1. 31 6 0. 05 4 0. 03 7 1. 46 9 0. 54 8 X L P 0. 03 5 0. 23 1 0. 15 3 0. 52 3 0. 05 5 9. 55 8* ** 0. 26 4 0. 06 4 4. 10 2* ** " 0. 21 9 0. 06 3 " 3. 50 4* ** 0. 05 0 0. 03 9 1. 27 7 0. 46 9 U U U 0. 15 8 0. 27 9 0. 56 5 0. 83 4 0. 06 6 12 .6 28 ** * " 0. 20 9 0. 07 8 " 2. 69 2* ** " 0. 10 7 0. 07 6 " 1. 41 7 0. 05 6 0. 04 8 1. 17 0 0. 57 8 X L U 0. 11 4 0. 31 3 0. 36 5 0. 59 5 0. 07 4 8. 02 9* ** 0. 38 2 0. 08 7 4. 38 6* ** " 0. 20 8 0. 08 5 " 2. 45 0* * 0. 08 1 0. 05 3 1. 51 4 0. 38 7 V V V 0. 14 5 0. 27 3 0. 53 0 0. 63 8 0. 06 5 9. 85 5* ** 0. 09 5 0. 07 6 1. 24 3 " 0. 01 5 0. 07 4 " 0. 20 7 0. 09 4 0. 04 7 2. 02 3* * 0. 41 5 X L V 0. 16 3 0. 22 8 0. 71 6 0. 71 2 0. 05 4 13 .1 81 ** * " 0. 06 5 0. 06 3 " 1. 03 0 " 0. 11 3 0. 06 2 " 1. 82 9 " 0. 02 3 0. 03 9 " 0. 59 2 0. 61 2 Y Y Y " 0. 17 6 0. 18 7 " 0. 94 1 0. 88 9 0. 04 4 19 .9 86 ** * 0. 31 6 0. 05 2 6. 05 3* ** 0. 24 8 0. 05 1 4. 87 3* ** " 0. 02 2 0. 03 2 " 0. 69 5 0. 79 9 X L Y 0. 06 3 0. 24 2 0. 26 2 1. 02 4 0. 05 7 17 .8 48 ** * 0. 28 0 0. 06 7 4. 15 0* ** 0. 16 8 0. 06 6 2. 56 8* * " 0. 10 6 0. 04 1 " 2. 56 9* * 0. 76 9 N ot e: B B B ,E E E ,F FF ,I II ,K K K ,P PP ,U U U ,V V V ,a nd Y Y Y re pr es en t th e Fi de lit y Se le ct Po rt fo lio m ut ua l fu nd s fo r th e M at er ia ls se ct or ,t he E ne rg y se ct or , th e Fi na nc ia l se ct or , th e In du st ri al s se ct or , th e T ec hn ol og y se ct or , th e C on su m er St ap le s se ct or , th e U til iti es se ct or , th e H ea lth ca re se ct or , an d th e C on su m er D is cr et io na ry se ct or ,r es pe ct iv el y. X L B ,X L E ,X L F, X L I, X L K ,X L P, X L U ,X L V ,a nd X L Y re pr es en t th e Se le ct Se ct or SP D R E T Fs fo r th es e ni ne se ct or s, re sp ec tiv el y. SP Y re pr es en ts th e SP D R S& P 50 0 T ru st . ** *S ig ni fic an t at th e 1% le ve l. ** Si gn ifi ca nt at th e 5% le ve l. 263C.Y. Lin / Financial Services Review 23 (2014) 249–271 model using peer ETF, respectively. On average, peer ETF benchmarking provides the best model fit for fund returns. This result is similar with Dellva, DeMaskey, and Smith (2001) and Kaushik, Pennathur, and Barnhart (2010). Table 8 One factor model results using peer ETF excess return, January 1999 through December 2010 Fund ! % Adjusted R2 Estimate SE t-value Estimate SE t-value BBB 0.506 0.196 2.588** 0.943 0.029 32.566*** 0.881 EEE 0.167 0.167 0.998 1.095 0.026 42.875*** 0.927 FFF 0.113 0.137 0.820 0.872 0.020 43.574*** 0.929 III 0.446 0.158 2.815*** 1.004 0.027 36.969*** 0.909 KKK 0.689 0.306 2.249** 1.170 0.038 30.712*** 0.868 PPP 0.332 0.175 1.895 0.831 0.047 17.529*** 0.681 UUU "0.064 0.316 "0.204 0.710 0.068 10.480*** 0.432 VVV 0.166 0.269 0.618 0.623 0.063 9.901*** 0.404 YYY "0.095 0.160 "0.595 0.764 0.027 28.011*** 0.845 Note: BBB, EEE, FFF, III, KKK, PPP, UUU, VVV, and YYY represent the Fidelity Select Portfolio mutual funds for the Materials sector, the Energy sector, the Financial sector, the Industrials sector, the Technology sector, the Consumer Staples sector, the Utilities sector, the Healthcare sector, and the Consumer Discretionary sector, respectively. ***Significant at the 1% level. **Significant at the 5% level. Table 9 Adjusted R2 comparison for different models, January 1999 through December 2010 Fund 1-Factor model (SPY) 3-Factor model (SPY) 4-Factor model (SPY) 1-Factor model (Peer ETF) BBB 0.590 0.656 0.655 0.881 XLB 0.620 0.672 0.670 EEE 0.363 0.379 0.389 0.927 XLE 0.352 0.378 0.394 FFF 0.644 0.783 0.786 0.929 XLF 0.638 0.796 0.802 III 0.789 0.856 0.855 0.909 XLI 0.778 0.820 0.821 KKK 0.617 0.901 0.900 0.868 XLK 0.715 0.894 0.899 PPP 0.363 0.544 0.548 0.681 XLP 0.315 0.467 0.469 UUU 0.557 0.577 0.578 0.432 XLU 0.240 0.381 0.387 VVV 0.406 0.402 0.415 0.404 XLV 0.609 0.614 0.612 YYY 0.737 0.800 0.799 0.845 XLY 0.728 0.760 0.769 Note: BBB, EEE, FFF, III, KKK, PPP, UUU, VVV, and YYY represent the Fidelity Select Portfolio mutual funds for the Materials sector, the Energy sector, the Financial sector, the Industrials sector, the Technology sector, the Consumer Staples sector, the Utilities sector, the Healthcare sector, and the Consumer Discretionary sector, respectively. XLB, XLE, XLF, XLI, XLK, XLP, XLU, XLV, and XLY represent the Select Sector SPDR ETFs for these nine sectors, respectively. SPY represents the SPDR S&P 500 Trust. 264 C.Y. Lin / Financial Services Review 23 (2014) 249–271 Comparing the Adjusted R2 of one-factor model with peer ETF to that of one-factor model with SPY, I find that seven out of nine MF regressions using peer ETF have a higher explanation power for the full and first half sample periods with the exceptions of the Utilities sector and the Healthcare sector. All nine MF regressions improve their Adjusted R2 for the second half sample period using peer ETF benchmark. Unlike previous research on sector mutual fund performance, this article is one of the first that provides detailed analysis on individual sector funds. Zheng and Tower (2005), for example, examine performance of asset-weighted and equal-weighted Fidelity sector fund portfolios and find that they performed less well than corresponding indexes. Dellva, DeMaskey, and Smith (2001) only list number of sector funds in their analysis. Kaushik, Pennathur, and Barnhart (2010) report sector aggregate performance results. For individual investors and/or their financial planners, however, it is important to examine individual sector fund performance and allocate their assets accordingly. 4.4. Changing dynamic Do MF and ETF in the same sector move together? Are they highly correlated with the general market index? Pairwise correlation is charted in Fig. 3. In all but two cases (Utilities and Healthcare) the correlation between MF and ETF in the same sector, the black bar, is the highest compared with the correlation between MF/ETF and SPY. Six sectors have a correlation between MF and ETF higher than 0.9 (Materials, Energy, Financial, Industrials, Technology, and Consumer Discretionary), whereas the other three sectors have a correlation of 0.827 (Consumer Staples), 0.661 (Utilities), and 0.638 (Healthcare). An interesting find was that these three sectors are often considered defensive sectors. Common factors in one sector appear to affect the returns of the MF and ETF in this sector more than that of general factors affecting the overall equity market. The returns of MFs and ETFs in the same sector tend to go hand-in-hand. The regression results during different sample periods discussed earlier, however, hint that 0.00 0.10 0.20 0.30 0.40 0.50 0.60 0.70 0.80 0.90 1.00 B E F I K P U V Y MF vs. ETF MF vs. SPY ETF vs. SPY Fig. 3. Monthly return correlation between funds, 1999–2010. B represents the Materials sector; E represents the Energy sector; F represents the Financial sector; I represents the Industrials sector; K represents the Technology sector; P represents the Consumer Staples sector; U represents the Utilities sector; V represents the Healthcare sector; Y represents the Consumer Discretionary sector; and SPY represents the SPDR S&P 500 Trust. 265C.Y. Lin / Financial Services Review 23 (2014) 249–271 the correlation might not be stable throughout the whole sample period. Fig. 4 plots 36-month rolling correlation for the nine sectors. Consistent with the results in Fig. 3, for most of the sample period, in all but two sectors (Utilities and Healthcare), the correlation between MF and ETF in the same sector, the black line, is the highest compared with the correlation between MF/ETF and SPY. The Industrial sector, the Technology sector, and the Consumer Discretionary sector have the most con- sistent correlation among three pairs of correlation plotted, MF versus ETF, MF versus SPY, and ETF versus SPY. The three lines are close to each other with the black one on the top. However, the Utilities sector and the Healthcare sector show great time varying correla- tion. The lines cross each other and the spread is huge. For example, for the Healthcare sector, the correlation between the ETF and SPY was the highest (0.864 vs. 0.127 and 0.189) for the first 36 months, January 1999 through December 2001, but it turns out to be the lowest (0.817 vs. 0.931 and 0.869) for the last 36 months, January 2008 through December 2010. A noticeable phenomenon is that the correlations tend to converge overtime. Almost all sectors have tighter correlation spreads moving into the end of the sample period. The correlation between the MF/ETF and SPY, the gray dashed line and the black dotted line, almost overlap for most of the sectors during the last quarter of the chart period. This changing dynamic explains why regression based results are sensitive to sample period selection. The open question is whether the correlation convergence will continue into the future. If yes, the diversification benefit one can enjoy through sector investing may diminish overtime. 5. Discussion 5.1. Benchmarking The Fidelity Select Portfolios cited both the S&P 500 and a MSCI U.S. IM sector 25/50 Index in the “Management’s Discussion of Fund Performance” section of the Annual Reports. However, the earliest ETFs based on the MSCI U.S. IM sector 25/50 Indexes were launched in January 2004 by Vanguard, which is five years later than the Select Sector SPDR Funds. The price correlation between these two ETFs in the same sector ranges from 0.975 to 0.999, with seven out of nine above 0.99, for the period from 2004 through 2010. This high correlation justifies the use of the Select Sector SPDR Funds as benchmarks for performance evaluation. One obvious explanation of the results that sector MFs outperform their peer ETFs somehow is that these actively managed mutual funds can invest outside of the S&P 500 basket. They can invest in foreign issuers as the prospectuses of the Fidelity Select Portfolios state. Domestically they can invest in any stock that is not in the S&P 500 index, mainly smaller capitalization stocks. They also only “normally … invest at least 80% of assets in securities of companies principally engaged in the selected sector,” which gives them some wiggle room across sector borders. However, it is hard to imagine they deviate very far from their benchmark. Only eight out 266 C.Y. Lin / Financial Services Review 23 (2014) 249–271 0.0 0.2 0.4 0.6 0.8 1.0 Ja n- 02 Ju n- 02 N ov -0 2 A pr -0 3 Se p- 03 Fe b- 04 Ju l- 04 D ec -0 4 M a y -0 5 O ct -0 5 M ar -0 6 A ug -0 6 Ja n- 07 Ju n- 07 N ov -0 7 A pr -0 8 Se p- 08 Fe b- 09 Ju l- 09 D ec -0 9 M ay -1 0 O ct -1 0 Materials 0.0 0.2 0.4 0.6 0.8 1.0 Ja n- 02 Ju n- 02 N ov -0 2 A pr -0 3 Se p- 03 Fe b- 04 Ju l- 04 D ec -0 4 M ay -0 5 O ct -0 5 M ar -0 6 A ug -0 6 Ja n- 07 Ju n- 07 N ov -0 7 A pr -0 8 Se p- 08 Fe b- 09 Ju l- 09 D ec -0 9 M ay -1 0 O ct -1 0 Energy 0.0 0.2 0.4 0.6 0.8 1.0 Ja n- 02 Ju n- 02 N ov -0 2 A pr -0 3 Se p- 03 Fe b- 04 Ju l- 04 D ec -0 4 M a y -0 5 O ct -0 5 M ar -0 6 A u g -0 6 Ja n- 07 Ju n- 07 N ov -0 7 A pr -0 8 Se p- 08 Fe b- 09 Ju l- 09 D ec -0 9 M ay -1 0 O ct -1 0 Financial 0.0 0.2 0.4 0.6 0.8 1.0 Ja n- 02 Ju n- 02 N ov -0 2 A pr -0 3 Se p- 03 Fe b- 04 Ju l- 04 D ec -0 4 M a y -0 5 O ct -0 5 M ar -0 6 A u g -0 6 Ja n- 07 Ju n- 07 N ov -0 7 A pr -0 8 Se p- 08 Fe b- 09 Ju l- 09 D ec -0 9 M ay -1 0 O ct -1 0 Industrials 0.0 0.2 0.4 0.6 0.8 1.0 Ja n- 02 Ju n- 02 N ov -0 2 A pr -0 3 Se p- 03 Fe b- 04 Ju l- 04 D ec -0 4 M ay -0 5 O ct -0 5 M ar -0 6 A u g -0 6 Ja n- 07 Ju n- 07 N ov -0 7 A pr -0 8 Se p- 08 Fe b- 09 Ju l- 09 D ec -0 9 M ay -1 0 O ct -1 0 Technology 0.0 0.2 0.4 0.6 0.8 1.0 Ja n- 02 Ju n- 02 N ov -0 2 A pr -0 3 Se p- 03 Fe b- 04 Ju l- 04 D ec -0 4 M ay -0 5 O ct -0 5 M ar -0 6 A ug -0 6 Ja n- 07 Ju n- 07 N ov -0 7 A pr -0 8 Se p- 08 Fe b- 09 Ju l- 09 D ec -0 9 M ay -1 0 O ct -1 0 Consumer Staples 0.0 0.2 0.4 0.6 0.8 1.0 Ja n- 02 Ju n- 02 N ov -0 2 A pr -0 3 Se p- 03 Fe b- 04 Ju l- 04 D ec -0 4 M ay -0 5 O ct -0 5 M ar -0 6 A ug -0 6 Ja n- 07 Ju n- 07 N ov -0 7 A pr -0 8 Se p- 08 Fe b- 09 Ju l- 09 D ec -0 9 M ay -1 0 O ct -1 0 Utilities 0.0 0.2 0.4 0.6 0.8 1.0 Ja n- 02 Ju n- 02 N ov -0 2 A pr -0 3 Se p- 03 Fe b- 04 Ju l- 04 D ec -0 4 M ay -0 5 O ct -0 5 M ar -0 6 A ug -0 6 Ja n- 07 Ju n- 07 N ov -0 7 A pr -0 8 Se p- 08 Fe b- 09 Ju l- 09 D ec -0 9 M ay -1 0 O ct -1 0 Healthcare 0.0 0.2 0.4 0.6 0.8 1.0 Ja n- 02 Ju n- 02 N ov -0 2 A pr -0 3 Se p- 03 Fe b- 04 Ju l- 04 D ec -0 4 M ay -0 5 O ct -0 5 M ar -0 6 A u g -0 6 Ja n- 07 Ju n- 07 N ov -0 7 A pr -0 8 Se p- 08 Fe b- 09 Ju l- 09 D ec -0 9 M a y -1 0 O ct -1 0 Consumer Discretionary MF vs. ETF MF vs. SPY ETF vs. SPY Fig. 4. Thirty-six months rolling correlation between funds. 267C.Y. Lin / Financial Services Review 23 (2014) 249–271 of 90 stocks of the top 10 holdings of the Fidelity Select Portfolios at the end of September 2012 are not in the S&P 500 index.14 Half of that eight are foreign stocks. Seven out of the eight has a weight between 1.59% and 4.05% of corresponding sector MFs; the other one weights 13.66%.15 The top weighted index stocks also make the backbone of the sector mutual funds. The number of stocks in the top 10 holdings of both sector MFs and ETFs varies from three to seven. Seven out of top 10 holdings of the Materials, Industrials, and Healthcare MF are also in the ETF top 10 list, and the top 10 holdings make up between 46% and 64% of these three mutual funds. That number is six for the Consumer Staples and Consumer Discretionary MF, five for the Energy and Technology MF, four for the Utilities MF, and three for the Financial MF. The top 10 holdings make up between 29% and 68% of the funds in these six sectors. 5.2. Fidelity Fidelity had been the largest mutual fund family that mainly provides actively managed mutual funds for several decades. Pozen and Hamacher (2011) argue that Fidelity, among several other fund families, has two characteristics that contribute to its success in the U.S. mutual fund business: dedication primary to asset management and control by investment professionals. The Fidelity fund family has maintained top market shares in the past decade: 10.2%, 11.8%, and 11.3% in year 1990, 1995, and 2010, respectively. It was No. 1 in 1990 and 1995, but passed by Vanguard (12.1%) in 2010. Fidelity’s stock is effectively controlled by members of its funding family, which relieves the short-term performance pressure from public shareholders to increase quarterly earnings. They also stated that Fidelity can develop compensation programs that promote top performance. Fidelity’s Megellan Fund has long been used as an evidence to defy EMH (see Kochman and Badarinathi 1993 and Marcus 1990). Fidelity’s large size can also potentially gain an insider edge as Golec (2007) shows some evidence on informed trades made by Fidelity funds. The large size of Fidelity’s assets under management can provide economy of scale for securities research, which aids the key element of active management: security selecting. Unlike passive managers, who often do little on stock selection, active managers can beat their benchmark by overweighting future winners, underweighting future losers, or some combination of both. They also have the freedom of holding cash. Focusing on only one sector, sector fund managers can potentially gain growing knowledge and experience dealing with stocks in that sector, which could help their performance. Elton, Gruber, and Green (2007) explain why funds may be more similar inside than outside fund families: Portfolio managers within families are likely to have access to the same research analysis produced either by internal analysts or by a particular set of external research firms; Portfolio managers may begin the security selection process with an eco- nomic forecast that is shared by other fund managers within the firm. It is not a surprise that Fidelity sector funds as a group can outperform when the shared research and macro view work well. 268 C.Y. Lin / Financial Services Review 23 (2014) 249–271 6. Conclusion This article is one of the first that reports detailed analysis on individual equity sector fund performance. It also contributes to the literature by adding evidence of active equity management outperformance. I find considerable evidence that sector mutual funds, the nine Fidelity Select Portfolios here, have provided better after-expense returns against broader market ETF, SPY, and their peer sector ETFs, the nine Select Sector SPDR Funds, during the sample period 1999–2010. Not only do they achieve higher nominal returns over the 12-year period except for few sector MFs, some of the funds also generate higher risk- adjusted returns measured by Sharpe Ratio16 and ! from various asset pricing models. None of the sector MFs generates a significant negative ! for the full sample period no matter which asset pricing model is used. That is, the sector MFs do not underperform SPY or peer ETFs measured by !. The Materials and the Technology sector MF stand out in the analysis. The Materials MF beats its peer ETF and SPY across the board for the full sample period: second highest 12-year return, highest Sharpe Ratio, and significant positive !s in all four regression models. The annualized abnormal return of the Materials sector MF is 11.856%, 8.628%, and 8.472% against SPY and 6.072% against XLB using factor models. The Technology MF also beats both SPY and peer ETFs on 12-year return, Sharpe Ratio, and most regression-based measures. It generates annualized abnormal return of 7.368% and 7.320% against SPY using the three-factor and four-factor models, and 8.268% against XLK. The Energy MF outper- forms on 12-year return, Sharpe Ratio, and one-factor model against SPY with an annualized abnormal return of 12.768%. The Industrials MF outperforms on 12-year return, Sharpe Ratio, and one-factor model against XLI with an annualized abnormal return of 5.352%. The Utilities and Consumer Discretionary MF have the weakest results. They do not beat their peer ETFs measured by 12-year return and Sharpe Ratio, however, they do not underperform when asset pricing models are adopted. Many researches have showed that on average active managers do not add value after fees and expenses (e.g., Barras, Scaillet, and Wermers 2010). Some argue that the value of active management lies in making market more efficient by improving asset allocation (Jones and Wermers 2011). Xiong et al. (2010) document that both asset allocation and active man- agement are critical to performance. The need for active managers to set the price was emphasized by a large index fund manager: “Passive management is a free-ride strategy; it piggybacks on active management. You need to have active managers out there, and they need to be paid.”17 This article, however, presents evidence on outperformance of equity sector funds. If one considers active managers together are playing a zero-sum game, this study finds some of the winners. For individual investors who are interested in sector investing, this study shows that actively managed mutual funds can be a good candidate for sector allocation. Most of these mutual funds do a better or equivalent job measured by both nominal return and total/ systematic risk adjusted return, after higher expenses and fees. The bottom line is that they do not underperform when measured with ! from asset pricing models, at least during the sample period. This study also provides evidence for financial planners when they help their 269C.Y. Lin / Financial Services Review 23 (2014) 249–271 clients select mutual funds. It may be appropriate for financial planners to recommend sector mutual funds to their clients who have risk appetite for sector investing. Notes 1 Most ETFs, although not all, are passively managed to track a specific index, such as the S&P 500. 2 https://www.fidelity.com/. 3 Because Telecommunication stocks are covered in the Technology Select Sector SPDR Fund, I do not consider Telecommunication a separate sector. 4 Fidelity Select Portfolio Annual Report, February 29, 2012. 5 Fidelity Select Portfolio Prospectuses. 6 However, all nine Select Sector SPDRs are diversified funds with respect to the Internal Revenue Code. As a result, each Sector Index is modified so that an individual security does not comprise more than 25% of the index. Source: http:// www.sectorspdr.com/. 7 Select Sector SPDRs Annual Report, September 30, 2011. 8 SPDR S&P 500 ETF Trust Annual Report, September 30, 2011. 9 The available price data for Fidelity Select Industrials Portfolio starts on July 8, 1999. 10 http://finance.yahoo.com/. 11 Subsample period results are discussed but not reported throughout the article. Results are available upon request. 12 For example, Ross, Westerfield, and Jordan, Corporate Finance, 9th ed., 2010, McGraw-Hill. 13 http://mba.tuck.dartmouth.edu/pages/faculty/ken.french/data_library.html. 14 From Fidelity and Sector Spider Web sites. 15 British American Tobacco PLC ADR represents 13.66% of the Fidelity Select Consumer Staples Portfolio as of September 28, 2012. 16 Eling (2008) shows that choosing a performance measure is not critical to fund evaluation and the Sharpe Ratio is generally adequate. 17 Frank J. Fabozzi, Sergio M. Focardi, and Caroline Jonas, 2010, Investment Manage- ment after Global Financial Crisis, Research Foundation Publications, CFA Institute, page 26. References Barras, L., Scaillet, O., & Wermers, R. (2010). False discoveries in mutual fund performance: measuring luck in estimated alphas. Journal of Finance, 65, 179–216. Carhart, M. M. (1997). On persistence in mutual fund performance. Journal of Finance, 52, 57–82. Dellva, W. L., DeMaskey, A. L. & Smith, C. A. (2001). Selectivity and market timing performance of fidelity sector mutual funds. Financial Review, 36, 39–54. Eling, M. (2008). Does the measure matter in the mutual fund industry? Financial Analysts Journal, 64, 54–66. Elton, E. J., Gruber, M. J. Green, C. (2007). The impact of mutual fund family membership on investor risk. Journal of Financial and Quantitative Analysis, 42, 257–277. 270 C.Y. Lin / Financial Services Review 23 (2014) 249–271 Fabozzi, F. J., Focardi, S. M. & Jonas, C. (2010). Investment management after the global financial crisis. Charlottesville, VA: Research Foundation Publications, CFA Institute. Fama, E. F., & French, K. R. (1993). Common risk factors in the returns on stocks and bonds. Journal of Financial Economics, 33, 3–53. Golec, J. (2007). Are the insider trades of a large institutional investor informed? Financial Review, 42, 161–190. Groysberg, B., Healy, P., Nohria, N., & Serafeim, G. (2011). What factors drive analyst forecasts? Financial Analysts Journal, 67, 18–29. Jones, R. C., & Wermers, R. (2011). Active management in mostly efficient markets. Financial Analysts Journal, 67, 29–45. Kacperczyk, M., Sialm, C. & Zheng, L. (2005). On the industry concentration of actively managed equity mutual funds. Journal of Finance, 60, 1983–2011. Kaushik, A., Pennathur, A. & Barnhart, S. (2010). Market timing and the determinants of performance of sector funds over the business cycle. Managerial Finance, 36, 583–602. Kochman, L. M., & Badarinathi, R. (1993). Net selectivity revisited. Journal of Economics and Finance, 17, 73–79. Marcus, A. J. (1990). The Magellan fund and market efficiency. Journal of Portfolio Management, 17, 85–88. McGahan, A. M. & Porter, M. E. (1997). How much does industry matter, really? Strategic Management Journal, 18, 15–30. Porter, M. E. (1985). competitive advantage: creating and sustaining superior performance. New York, NY: Free Press. Pozen, R., & Hamacher, T. (2011). Most likely to succeed: Leadership in the fund industry. Financial Analysts Journal, 67, 21–28. Sullivan, R. N., & Xiong, J. X. (2012). How index trading increases market vulnerability. Financial Analysts Journal, 68, 70–84. Xiong, J. X., Ibbotson, R. G., Idzorek, T. M., & Chen, P. (2010). The equal importance of asset allocation and active management. Financial Analysts Journal, 66, 22–30. Zheng, W., & Tower, E. (2005). Fidelity versus vanguard: Comparing the performance of the two largest mutual fund families. International Review of Economics and Business, 52, 433–465. 271C.Y. Lin / Financial Services Review 23 (2014) 249–271