Vanguard bond index fund tracking error

From Bogleheads
Jump to: navigation, search


Vanguard bond index fund tracking error measures Vanguard's indexing performance. Tracking error is the ultimate measure of judging an index fund manager's performance. Since bond indexes can include huge numbers of illiquid bonds, replicating an index is often very costly. Vanguard, in an attempt to control costs, samples securities in their bond index funds. The sampling attempts to match the sector and risk metrics of the underlying bonds in the index. Because a sampled index fund does not hold all of the securities in the underlying index, its returns may vary somewhat from those of the index. Such performance variance is termed "sampling error." Over time, Vanguard has made a number of changes in the sampling policy used by the funds. In 2009, the funds adopted a change in benchmark tracking indices, adopting the Barclays Capital U.S Government/Credit Float Adjusted Indices. [1]

Tracking error

Table 1. provides the long term average tracking errors for Vanguard's bond index funds. Data extends from 1995 to present for the Short, Intermediate, and Long term funds (from the first full year of operations). Data extends from 1993 to present for the Total Bond Index (first full year 1987). Admiral share class data covers the period from 2002 (the first full year of operations of the share class) to the present. Detailed annual tracking error data, along with additional statistical measurement of fund returns, are provided in the footnote tables.


Table 1. Vanguard Index Fund Tracking Errors (compound total return) [2]
Fund Investor shares Admiral shares
Year Tracking Error Year Tracking Error
Vanguard Short Term Bond Index [note 1]
Investor Admiral
1995-2014 -0.20% 2002-2014 -0.19%
Vanguard Intermediate Term Bond Index [note 2] Investor Admiral 1995-2014 -0.15% 2002-2014 -0.14%
Vanguard Total Bond Market Index [note 3]
Investor Admiral
1993-2014 -0.21% 2002-2014 -0.18%
Vanguard Long Term Bond Index [note 4]
Investor
1995-2014 -0.08% n/a n/a

Sampling strategies

From inception (1987 and 1995) the Vanguard bond index funds have employed sampling strategies in implementing investment policy. In addition, the funds originally employed a corporate bond substitution policy, whereby they substituted short term corporate bonds for treasury bonds in fund holdings. The funds' substitutions were limited to securities with less than 4 years remaining to maturity, and were restricted to a maximum of 15% of total assets. The rationale for substitution was that the higher interest income from corporate issues would help the funds offset the costs of implementation and would, with a moderate increase in diversifiable risk, allow the funds to more closely track their benchmark indices. [note 5]

As shown in the second column of Table 2. below, the substitution policy resulted in the funds providing low tracking errors over the 1995-2001 period. In every instance, the funds provided returns that partially offset the drag of expenses. In 2002, a substantial increase in market risk and defaults hit the corporate bond market. As a result of the corporate substitution policy, the Vanguard bond index funds failed to closely track their benchmark indices. The funds held higher than benchmark weightings of telecommunication and energy company bonds, both of which were hurt by defaults and credit quality erosion. [3] Column three, in Table 2. shows the fund's tracking errors in 2002. The short term, intermediate term, and total bond portfolios all lagged benchmark returns by over 2.00% for the year. Column four shows the tracking error of the funds over the 1995-2002 period.

Vanguard initially responded to the performance lag by reducing the allowable corporate substitution in the funds to a maximum 10% of net assets. By early 2005 [4] the funds adopted a sampling strategy that required the funds to more tightly sample not only industry sectors, but also to tightly sample industry subsectors. In 2009, as a result of the 2008-2009 financial crisis and the expansion of Federal Reserve purchasing of mortgage backed and other debt market securities, the Vanguard bond funds adopted free float versions of the Barclay bond indices as benchmark indices. [note 6] The fifth column in Table 2. shows the tracking errors realized by the funds subsequent to the 2002 year and the changed sampling policy.

In addition, Table 2. (columns six and seven) show the tracking errors for the funds' investor shares and admiral shares over a common period (from inception of admiral shares in 2002 to present). The lower expense admiral shares can more closely track the benchmark.

Table 2. Vanguard Index Fund Tracking Errors
Fund Investor Tracking Error Investor Tracking Error
(2002-2014)
Admiral Tracking Error
(2002-2014)
1995-2001 2002 1995-2002 2003-2014
Vanguard Short Term Bond Index -0.06% -2.02% -0.31% -0.13% -0.28% -0.19%
Vanguard Intermediate Bond Index -0.02% -2.18% -0.26% -0.08% -0.24% -0.15%
Vanguard Total Bond Market Index -0.07% -2.00% -0.27% -0.14% -0.29% -0.19%
Vanguard Long Term Bond Index 0.08% -0.46% 0.01% -0.13% n/a n/a

Notes

  1. Table 3.
    Vanguard Short Term Bond Index Fund

    (View Google Spreadsheet in browser, then File --> Download as to download the file.)
  2. Table 4.
    Vanguard Intermediate Term Bond Index Fund

    (View Google Spreadsheet in browser, then File --> Download as to download the file.)
  3. Table 5.
    Vanguard Total Bond Market Index Fund

    (View Google Spreadsheet in browser, then File --> Download as to download the file.)
  4. Table 6.
    Vanguard Long Term Bond Index Fund


    (View Google Spreadsheet in browser, then File --> Download as to download the file.)
  5. SEC 485 1995 prospectus:
    The Portfolios of the Fund may, from time to time, substitute one type of investment grade bond for another. For instance, a Portfolio may hold more short-term corporate bonds (fewer short U.S. Treasury bonds) than represented in the Index so as to increase income. This corporate substitution strategy will entail the assumption of additional credit risk; however, substantial diversification within the corporate sector should moderate issue-specific credit risk. In addition, current investment policy restricts corporate substitutions to issues with less than 4 years remaining to maturity and in aggregate no more than 15% of net assets. Overall, credit risk is expected to be very low for each of the Portfolios.
  6. SEC 485 2009 prospectus:
    Index sampling strategy. Because it would be very expensive and inefficient to buy and sell all bonds held in its target index—which is an indexing strategy called “replication”— the Fund uses index “sampling” techniques to select securities. Using sophisticated computer programs, the Fund’s advisor generally selects a representative sample of securities that approximates the full target index in terms of key risk factors and other characteristics. These factors include duration, cash flow, quality, and callability of the underlying bonds. In addition, the Fund keeps industry sector and subsector exposure within tight boundaries relative to its target index. Because the Fund does not hold all issues in its target indexes, some of the issues (and issuers) that are held will likely be overweighted (or underweighted) compared with the target index. The maximum overweight (or underweight) is constrained at the issuer level with the goal of producing well-diversified credit exposure in the portfolio.

    A float-adjusted index is an index that weights its constituent securities based on the value of the constituent securities that are available for public trading, rather than the value of all constituent securities. Some portion of an issuer's securities may be unavailable for public trading because, for example, those securities are owned by company insiders on a restricted basis or by a government agency. By excluding unavailable securities, float-adjusted indexes can produce a more accurate picture of the returns actually experienced by investors in the measured market.

References