So How Big is the Indexing Market in the United States?

Two months ago, we asked ourselves what we thought was this relatively simple question. We now know that the answer to it is anything but simple, or easy to determine. At the end of this article, we'll offer our opinion as to why this is the case. But first, we'll answer the question that we asked.

The first challenge you face when attempting to discover the size of the indexing market in the United States is collecting the data you need. Suffice it to say, it doesn't exist in one location, and is far from neatly organized. Specifically, we found that the raw data we had to work with was divided between information about the type of investors who indexed their funds, and the type of instruments they used.

Broadly speaking, there are two general types of investor: individuals, and institutions such as pension plans, bank trust departments, and life insurance companies. To index their investments, these investors use a variety of instruments, including index mutual funds, and the direct holding of the bonds or equities (including exchange traded index funds) that comprise the index whose performance they are trying to track.

Given this, to estimate the size of the U.S. indexing market, we have tried to fill in all the cells in a 2 x 2 matrix that looks like this:

Let's start with the upper right cell, direct holdings by institutional investors. Every year, Pensions and Investments Magazine surveys the top 1000 pension plans in the United States, ranked by the size of the assets they have under management. Broadly speaking, pension plans fall into two main categories. Defined benefit plans pool all employees' retirement savings, and in return guarantee a certain level of retirement income. These savings are allocated to different asset classes, and investments within each asset class are generally managed by professional investment managers, under the supervision of the plan trustee. Defined benefit plans place quite a bit of risk on the shoulder's of the companies that sponsor them. If their investment performance does not generate sufficient funds to pay the anticipated benefits to plan participants, the company has to increase the size of the contributions it makes to the plan, which reduces the funds it has available to distribute to its shareholders, and/or invest in new projects. Because of this, defined benefit plans have increasingly been replaced by defined contribution plans, where the company only guarantees that it will contribute certain amounts to the plan, but does not guarantee a minimum level of income to retirees. Moreover, under defined contribution plans (which are often known by the chapter of the regulations they are structured under, such as 401 (k) or 403 (b)), employees themselves (rather than professional money managers) are largely responsible for deciding how to invest their retirement savings.

At the end of 2001 (the year we will use for all our estimates), the top 1000 retirement plans were estimated to have $3.6 trillion in defined benefit assets. A survey of the top 200 plans, which represent approximately 75 percent of the assets controlled by the top 1000 plans, showed that 30% of their assets were explicitly invested using an indexing approach. We say "explicitly", because their has been much discussion in recent years of "de jure" versus "de facto" (also known as "closet") indexing. The latter case is when an active manager, to avoid losing an investment management account, structures his or her portfolio to closely track the benchmark index that is used to measure his or her performance. As evidence of the growth of "closet indexing", analysts frequently cite the sharp decline in active funds' "tracking errors" versus their performance benchmarks in recent years. Given that there is no way to estimate the extent of "closet indexing" (except by a wild guess), we have elected to include in our estimate only those funds that are explicitly indexed. However, given the widespread practice of "closet indexing" our estimate should be regarded as on the low side of what is probably the case in practice. That being said, to calculate our estimate, we assumed that the 30% indexing figure for the top 200 defined benefit plans also applied to the other 800 plans. Given that the total assets under management at the top 1000 defined benefit plans amounted to $3.6 trillion at the end of 2001, approximately $1.08 trillion (30% times $3.6 trillion) in assets were managed using an index approach. We should also note that our 30% estimate falls roughly in the middle of two similar estimates that were recently made. Goldman Sachs also studied a sample of 200 top defined benefit plans, and estimated that 25% of their assets were indexed. At the other end of the spectrum, the Sandler Review published by the U.K. Treasury in July, 2002 estimated that 40% of U.S. institutional assets were indexed.

The second type of institutional investor with significant direct security holdings are life insurance companies. According to the Federal Reserve's Flow of Funds data, at the end of 2001, life companies directly held $855 billion in corporate equities, and $2,075 billion (that is, $2.1 trillion) in credit market instruments (government and corporate bonds, mortgages, and loans). We could find no estimate about how, in aggregate, these funds are managed. Given that, we elected to focus on a subset of the life industry where data was available: the annuity market. In recent years, variable annuities have become quite popular. In simplified terms, they are a bundle of mutual funds (called subaccounts), wrapped up in an insurance product (which has some tax advantages). As in the mutual fund business, life companies have also found some of their annuity customers asking for indexed products. They have responded by offering index subaccounts, and products known as indexed annuities. In 2001, the total value of variable annuity sales was $113 billion. Of this amount, approximately $7 billion (or 6.2%) were indexed annuities. At the end of the year, the value of annuity assets (that is, the value of the funds received from the sale of annuity products in previous years, plus the appreciation of those funds) was $883 billion. We estimated that 6.2% of this (or about $55 billion) represented indexed annuity assets. Given that indexed annuities have been rapidly rising in popularity only in recent years, this estimate may be on the high side; however, given the growing relative market share of indexed products, we don't think it is too far off the mark. (We should also note that we have not included in this calculation the value of assets supporting the approximately $60 million in premiums that are received each year from the sale of equity indexed life insurance, as opposed to annuity products.).

The third major type of institutional investor is bank trust departments. Federal Reserve Flow of Funds data showed that they directly held $33 billion in corporate equities and $229 billion in credit market instruments at the end of 2001. Our sense is that, like life insurance companies, bank trust departments are more likely to manage their investment portfolios internally than they are to use outside money managers. Given this, we assume that they are also more likely to employ active management. For this reason, we have used the lower estimate of 6.2% of assets managed using indexing. Given directly owned securities of $262 billion, this yields an estimate of $16 billion in indexed assets.

Institutional investors not only own securities directly, they also own mutual funds. The largest institutional owners of mutual funds are direct contribution retirement plans. Pensions and Investments estimated that the top 1000 retirement plans owned $1.2 billion in defined contribution assets at the end of 2001. The top 200 defined contribution plans had assets of about $706 billion. Of this amount, approximately 55 percent ($385 billion) was in corporate 401(k) plans. On average, these corporate plans had 32% of their assets invested in their own company's stock, leaving about $262 billion ($385 x 68%) for other investments. The remaining defined contribution plans in the top 200, with assets of $321 billion, were sponsored by public sector entities and unions, and did not invest in company stock. Of the total amount of defined contribution assets not invested in company stock ($262 +$321 = $583, or 83% of $706), 17% was indexed.

At the end of 2001, the Investment Company Institute (the main trade association for mutual funds) said that it recorded $1,200 billion in mutual fund assets that were held by defined contribution plans. Assuming 17% of these represented investments in index funds, the total amount of indexed defined contribution assets was $204 billion. At the same time, bank trust departments were reported (in Federal Reserve Flow of Funds data) as holding $359 billion in mutual fund assets. Assuming the same previously estimated 6.2% indexing to total assets ratio applies to these holdings, we estimate that bank trust departments held a further $22 billion in indexed assets.

The Investment Company Institute also reported that, at the end of 2001, of $4,689 billion in long term (basically debt and equity, but not money market) mutual fund assets, $366 billion was invested in index mutual funds. Given that we have estimate that $226 billion in indexed mutual fund assets were held by institutional investors, this leaves $140 billion in indexed mutual fund assets that were held by individuals, in either taxable or tax-exempt (IRA) accounts. We also know that of the $4,689 in long term mutual fund assets, $1,200 billion was held by direct contribution pension plans, $359 billion was held by bank trust departments, $108 billion was held by non-financial corporations, and $44 billion was held by life insurance companies. This leaves $2,978 in mutual fund assets that were held by individuals. Our estimated indexed assets of $140 billion therefore represents only 4.7% of the mutual fund assets held by individuals.

This brings us to individuals' direct holdings of debt and equity securities (the lower right hand box in our matrix). While individuals directly owned substantial amounts of debt and equity securities at the end of 2001 (respectively, $2.4 billion and $6.0 billion, according to the Federal Reserve Flow of Funds data), we think it is highly unlikely that much of this money was indexed. There are, however, two exceptions to this. The first is the $416 billion that was held in separately managed accounts (basically, direct holdings managed by a money manager hired by the owner of the securities). Because virtually all of these assets were owned by high net worth individuals (due to the high minimum investments required to establish a separate account), and because these investors tend to be more sophisticated, we have estimated that 25% of these funds were indexed, or $104 billion. Our basis for this estimate was the Sandler Report's finding that "for funds sold direct to consumers, the proportion of passive management was 25%, which is essentially the same proportion as in the [U.K.] institutional world. This is consistent with a situation in which the more knowledgeable consumers, who feel able to buy direct, are aware of the difficulty of identifying superior managers", and are thus more likely to prefer indexing.

To this amount, we also need to add the assets of exchange traded index funds that are directly owned by individuals. At the end of 2001, ETF assets amounted to $83 billion. However, while we could not find an exact estimate, there were multiple reports that institutional investors were by far the heaviest users of ETFs. We have assumed that individuals held only 30% of ETF assets at the end of 2001, or about $25 billion. To avoid double counting, we assume that the remaining ETF assets were included in the direct holdings of institutional investors, and separately managed individual accounts.

We have now completely filled in our matrix. Here is what it looks like:

Three conclusions immediately emerge from this matrix: first, we estimate that at least $1.6 trillion in assets were indexed in the United States at the end of 2001. Second, most indexed assets are held outside of mutual funds. Finally, institutional investors are, relative to the assets they manage, much bigger users of indexing than are individual investors. This last observation brings us full circle to the question we asked at the beginning of this article: why was it so hard for us to get an accurate picture of the extent of index investing in the United States?

The Sandler Review raised some interesting points in this regard. It noted that "one of the key decisions an investor must make is whether his or her investments should be managed actively or passively [that is, indexed]. Active management is only appropriate where the investor believes that (a) there are systematic inefficiencies in the market that can be exploited; (b) he or she can identify in advance those investment managers capable of doing so; and (c) the potential benefits outweigh the additional costs involved [beyond those charged by passive managers]…One would [therefore] logically expect to find that institutional investors as a group made greater use of active management than individual investors, since institutions typically have greater resources and professionalism to bear in the investigation and selection of active managers. In fact, the reverse is the case…Large numbers of individual investors, who are not in a position to identify superior managers [in advance] are nonetheless investing in actively managed funds…Data from a mystery shopping exercise analyzed by the Review suggested that consumers' preference for active funds is shared by their advisers [note that in the United States, only 15% of mutual funds are sold direct to individual investors]. Ninety two percent of advisers' fund recommendations were for active funds. Subsequent interviews with advisers confirmed the existence of this preference." Then in a great bit of British understatement, the Review concluded that "it would be implausible to attribute this preference for active management, which is materially greater than that demonstrated by institutional investors, to superior expertise on the part of the advisers to individual investors."

We think the challenges we faced in putting together this data on indexing reinforces the Sandler Review's conclusion that the investment management industry, and those who sell its products to individual investors, have a very strong disincentive to fully informing the investing public about the benefits of indexing. When you are earning high incomes from sales commissions and high management fees on actively managed products, it is not in your self-interest to tell your customers about a different approach that costs less and delivers better results over the long term. Critically, undoubtedly includes a resistance to showing individual investors the extent to which the most sophisticated institutional investors have chosen to index their investments. Sadly, we think that is why our task has been so difficult.

On the other hand, now that it is done, we hope that the estimates we have developed will inspire more individual investors to reconsider their use of active management, and focus instead on the long-term benefits of index investing.

| So Where Does This Leave Us? | Equity Market Valuation Update | So How Big is the Indexing Market in the United States? | Historical Asset Class Return | Why Asset Allocation is Still A Challenge | Model Portfolio Update | Future Class Returns |