Pig in the Poke? Why It’s Time to Bag the Agg

The first half of 2023 revealed the Agg drag. For the three-year period ending 1Q 2023, the Agg’s not so Sharpe ratio was -0.56%. As for diversification, the top two components of the Agg had an 81% correlation to one another. In July 2023, the Agg’s effective duration (6.3 years) remained elevated (47% above its average) versus its historical average of 4.5 years since the measure’s inception in 1976.⁴ This seemingly innocuous 1.8-year extension in duration would result in additional losses of roughly 1.8% if bond yields were to rise by 100 basis points. The Agg’s standard deviation was 6.19%. These

The origins and ubiquity of financial indices

Financial indices like the Dow Jones Industrial Average® (DJIA®), the Consumer Price Index (CPI), the S&P 500® Index (SPX), the Bloomberg US Aggregate Bond Index (the Agg), and the VIX® Index are the indispensable structures of modern finance. These indexes measure the daily pulse of financial markets – the backbone of investments and finance.

Indices are used as blueprints for investments, referents for contracts, and as sources of information.⁵ There are three types of financial indices –­ public, product and byproduct indexes. (The terms “indices” and “indexes” are used interchangeably here.) There are tens of trillions of assets invested in vehicles that track indices and hundreds of trillions of dollars contracted based on an index reference.

One of the most important uses of financial indices is as investment blueprints (benchmarks). In the early 20th century, investment managers served as both architects and general contractors for actively managed strategies tied to financial indexes and benchmarks. Benchmark indices are used by financial advisors, institutional and retail investors for three primary reasons: 1) as portfolio performance targets; 2) as informed measures of security-level and asset class risk and return traits; and 3) as references for index-linked products.⁶

The Dow Jones Industrial Average®, which debuted in May 1896, is regularly cited as the granddaddy of all benchmarks. However, the DJIA® was actually launched 12 years after its sister index, the Dow Jones Transportation Average, which first published on July 3, 1884.⁷ The U.S. Senate got into the index act in 1902 when it resolved to create the first federal index – a progenitor of the Producer Price Index (PPI).⁸ The S&P 500® Index was created in 1957 and the first total-return bond index in 1973.⁹ The VIX® Index, created by Professor Robert Whaley, was launched by the CBOE on January 19, 1993.

Today, the sheer number of financial indices and benchmarks is astonishing. At the Index Industry Association’s last survey (2018), there were 3.288 million financial indices worldwide.¹⁰ This Cambrian index explosion covers every possible asset class, domestic and international, investment-grade and high-yield credit, currencies, ETFs and ETPs, volatility and inverse, sin stocks or ESG, smart beta and dumb alpha, and customized or bespoke benchmarks.¹¹ When this survey of global indices was conducted, there were 70 times more stock market indexes than listed public companies.¹² There were also over 525,000 fixed income indexes or benchmarks.

The arbitrary anatomy of financial indexes

Providing an index is a goal-oriented activity. Financial indexes can be used across all stages of the investment process – including conception, construction, benchmarking versus peers and performance attribution. Indices enable cheaper forms of traditional investing and new forms of investable assets by facilitating derivative contracts. Many derivative or futures contracts use a financial index value for settlement to determine who owes whom how much.¹³

Financial indices also convey information and increase price transparency. Indices are irreplaceable because they solve collective action problems implicit in price discovery. Research costs can become duplicative and wasteful if competing parties are seeking an edge. A public index, like the CPI, or an equity index (e.g., S&P 500®) can mitigate excessive price research by allowing interested parties to share the results of price discovery and its costs.

There are numerous roles of a benchmark or a financial index. An index should effectively represent “the market” by delivering an unbiased, complete view as an investment blueprint of the segment it is designed to track. The method by which the index constituents are selected should be free of subjectivity.

Before continuing to extoll the benefits of financial indexes, it is essential to divulge an index’s Achilles heel. The construction and maintenance of virtually every financial index is fundamentally discretionary, subjective, and even capricious. Financial indices may provide recipes for diversified asset class exposures, but the index ingredients are determined by advisory committees and/or individuals who have biases.

Some index providers deliberately reduce the accuracy of their index. All indices – bond or stock – exist on a benchmarking spectrum where accuracy is balanced against other legitimate goals, including tradability and consistency.¹⁴ Subjective judgments are an integral part of indices’ construction and maintenance. This subjectivity also carries the risk of index manipulation (LIBOR Index was index manipulation’s Exhibit A), front-running and malproduction, which includes the failure to consider changes in underlying components of the index.

(The now discredited and defunct LIBOR Index was calculated and set every day by John Ewan, a “junior” employee at the British Bankers Association (BBA). Ewan was described as “significantly out of his depth.” In a 2008 email, then BBA CEO, Angela Knight, wrote, “Never again can the LIBOR Desk be the place where a ‘junior’ gets put to see if they are going to make the grade”. John Ewan, at the time in his mid-30s, stayed on for another four years.¹⁵)

Fixed-income benchmarks encounter three significant drawbacks: a) the duration problem; b) the “bums” problem; and c) sampling. The duration structure of a cap-weighted bond benchmark – i.e., the proportions of bonds in short-, intermediate-, and long-term categories – arises from the maturity or duration preferences of issuers, who seek to minimize their cost of capital. Investors, on the other hand, aren’t trying to minimize returns; they want to maximize risk-adjusted returns. The benchmark duration of the Agg is a historical accident. The optimal portfolio for an investor with no defined time horizon should be set by that investor’s risk tolerance rather than by matching the duration preferences of bond issuers.

Because traditional fixed-income benchmarks usually weight securities based on market capitalization, there’s a bias to highly indebted countries and firms. The “bums” (or deadbeats) problem is that the biggest debtors (whether companies, countries, or other entities) have the largest weights in the benchmark. Issuers who go deepest into debt – the biggest bums – have the largest weights in a cap-weighted benchmark. Cap-weighted indexes must buy bonds in proportion to their capitalization weight to minimize tracking error to the benchmark, even if the security is only marginally of high enough quality to make it into the benchmark. These securities are the most likely to be downgraded or to default.¹⁶ This has a highly undesirable effect. Issuers that can pay back their debts should be more in demand.

The third bond benchmark bummer is the sampling problem. Frequently, stock-index trackers use an exact-replication strategy, whereas bond index trackers use a sampling approach. But even the most exact replication of an equity index has additional filters, all subject to human discretion. For example, the S&P 500 imposes profitability and domicile requirements, but its selection committee waives them on a case-by-case basis for popular or important firms.

Due to the sheer number of bonds in Bloomberg’s Bond Indices (13,368 in the U.S. Agg; 25,000+ in the Global Aggregate Index), passive strategies often use sampling (representative sample of bonds that meet the risk profile and characteristics of the parent benchmark). Sampled baskets lead to a reduced security count relative to the index and force passive indexers to skew their allocations to larger and more liquid issues. Many passive bond strategies ignore smaller- to medium-sized debt securities or healthier corporations who don’t need or want to issue a lot of debt.

Since bond indexes can include huge numbers of illiquid bonds, replicating an index is often very costly. To control transaction costs, sponsors of bond-index-tracking funds and ETFs hold only a fraction of the securities in the index. Bond indexes and funds that track them are designed to match the index with respect to sector allocation, duration, cash flows, quality and callability.¹⁷ Because a sampled index fund does not hold all of the securities in the underlying index, its returns may vary from those of the index itself.

The typical rule for an index provider is to have 80% of their assets invested in bonds in the index and the remaining 20% invested outside the index. The outside investments are usually in more liquid instruments, such as futures, options, and interest rate swaps, but also could be in nonpublic bonds or lower-rated bonds.¹⁸ The result of this trifecta of drawbacks means investing in a broad bond benchmark can markedly increase the level of interest-rate and credit risk exposures without investors realizing it.

Index providers must balance conflicting objectives – accuracy versus tradability versus index consistency – and make value judgments about index composition. No index can maximize every value, and index stewards must make tradeoffs among at least these three variables.

The index advisory committee or “chief indexer” incorporates subjective intervention and judgment into an index’s creation and formation, maintenance, and rebalancing. There is no unbiased or perfect index. Adam Smith’s “invisible hand” is a very human appendage in the world of three million-plus indices and benchmarks.

Aggregate Bond Index Excludes Large Parts of U.S. Bond Market

Disaggregating the Agg

“Bond selection is primarily a negative art. It is a process of exclusion and rejection, rather than of search and acceptance.” ― Benjamin Graham, Security Analysis: Principles and Technique

The Agg is a market-cap-weighted, investment-grade (must be Baa3/BBB- or higher), U.S. dollar-denominated, intermediate, fixed-rate coupon, taxable bond benchmark. It is composed almost exclusively of U.S. Treasury and agency mortgage-backed securities (roughly 70%), U.S. dollar-denominated sovereign bonds, high-quality corporate bonds (or “credit”) and a small number of securitized instruments. “Intermediate” means there must be at least one-year to final maturity.

The Agg’s rules-based construction intentionally excludes many fixed income securities that investors prefer in a modern, well-diversified portfolio. While the “aggregate” name implies comprehensiveness, Bloomberg makes it clear that the Agg isn’t designed to provide exposure to the entirety of the fixed-income universe. By design, the Agg omits high-yield credit, floating-rate bonds, inflation-linked bonds (TIPS), tax-exempt municipals, convertible bonds and structured notes, certain agency mortgage securities and most asset-backed securities.

The Agg is discretionary and exclusionary. The Bloomberg Index Services Limited (BISL) publishes a wide range of bond and equity indices. The Bloomberg Fixed Income Indices (US Aggregate Indices) are governed by two primary internal committees and an external Index Advisory Council. For example, in 2017, the US Aggregate Indices methodology changed, and new restrictions were enacted:

• For Treasury, government-related and corporate securities in the U.S. Aggregate Index, the minimum amount outstanding was raised from $250 million to $300 million. This resulted in 1,023 securities dropping from the U.S. Agg, equivalent to $304 billion in market value.
• For MBS pass-throughs, pools must have USD $1 billion in par amount outstanding.
• MBS Hybrid ARMs were removed from the U.S. Aggregate starting June 1, 2017.

The Agg favors the largest, most liquid bond issuers. For Treasury, government, and corporate securities, there is a floor or minimum of $300 million par outstanding for membership in the Agg. This results in a perverse dynamic in the Agg where companies, countries and issuers who issue more debt and become less creditworthy are allocated a bigger share of the index.

J.P. Morgan research claims that the Agg captures only 49% of the US bond market – a woefully incomplete picture.¹⁹ The Agg has $25 trillion in market capitalization.

While the Agg may have historically been viewed as a core and stable component of diversified portfolios, it deserves the same scrutiny as other indices because it has changed. There are crucial litmus tests where the Agg malfunctions. If an index is supposed to deliver an unbiased, complete view of the market or market segment it is designed to track, the Agg fails miserably in this regard.

Defragging the Agg

If the DJIA® is the granddaddy of equity indexes, the Agg is the grandma of bond indexes. The world’s oldest total-return bond index traces its lineage back to the 1970s. This aggregate bond index concept was conceived by two Kuhn Loeb researchers in July 1973. At the time, bond indices consisting of yield averages had been around for decades, but a fixed income total-return benchmark did not exist.²⁰ The Agg had historically backfilled to January 1, 1976. After a series of acquisitions culminating in 2016, it became known as the Bloomberg US Aggregate Bond Index.

Last year was the worst on record for the Agg – losing 13.01% for the year (see image above). With data going back to 1973, the Agg had two consecutive calendar years of negative returns for the first time ever in 2021-22. Before 2021, the index had only been negative four times in 46 years.²¹ The Agg was down 1.54% in calendar 2021. The Agg had an annualized return of negative 0.10% over the past five years; 2022 was the first calendar year ever to end with a negative five-year annualized return. U.S. and global bonds markets were in the grasp of a raging bond bear for the first time in a generation.

A regime change is unfolding, with rising interest rates and bond yields and persistently stubborn inflation. It’s instructive to consider real (inflation-adjusted) returns in last year’s bond rout. The Agg lost 13.01% and the annualized CPI rose 6.5% in 2022.²² In real terms, passive intermediate investment-grade bond index investors lost 19.6%.

In June 2023, the Agg was composed of 13,368 bonds worth $25.1 trillion with total U.S. government exposure of roughly 70%.²³ Bonds constantly exit (and enter) the Agg for three main reasons: 1) time to maturity becomes less than one year; 2) issuers call (redeem) their bonds; and 3) rating agencies downgrade a bond from investment-grade to speculative grade (a fallen angel) or upgrade a bond from junk status to IG. And yearly new issuances – particularly of government paper – are a main reason new bonds enter the Agg. This results in a much higher turnover in big bond indexes versus stock index peers like the Dow Jones, MSCI or S&P 500 indexes.

Despite 13,000+ bonds, 95+% of the Agg is composed of only four types of debt: U.S. government, government-related, agency mortgage, and high-quality corporate debt. The top two components of the Agg have an 81% correlation to one another. This limits the Agg Index’s sources of returns to cash and returns derived from three factors: interest-rate exposure, investment-grade corporate spreads over comparable maturity Treasury bonds and mortgage spreads. A market value-weighted benchmark of investment-grade (IG) debt is overwhelmingly influenced by the performance of a single dominant risk factor – interest rate risk.

There’s one more drag to the Agg. Besides the limited number of holdings relative to their benchmarks, passive bond funds suffer from an affliction of “zero-trading days.” About one half of assets in the benchmark of the average passive bond fund do not trade in a given day. A recent survey of the distribution of zero-trading days of the Agg discovered at the 90th percentile, bonds in the Agg trade only one out of every seven trading days. The other 10% of the Agg’s bond issues trade even less frequently.²⁴ The opaque nature of bond markets causes havoc for price discovery.

Given the size of the debt markets, there’s ample diversification potential in other bond sectors.

In the 1980s, American Express’ advertising slogan was, “Membership has its privileges”. The Agg severely limits its members. Some of the best fixed income sectors today and historically have been in opportunistic bond markets ineligible for membership in the Agg. Non-agency mortgage-backed securities (MBS), emerging market local debt, high-yield (HY or junk bonds), and non-dollar denominated are ineligible for the Agg.

Unshackle your fixed income opportunity set

The key takeaway from this expose? Don’t invest in a passively managed bond index fund unless you know its strengths and shortcomings. Given the difficulty of replicating the Agg’s characteristics, and the risks inherent to passive bond strategies, investors should shun the Agg Index because of its limitations. A market value-weighted benchmark of investment-grade bonds is overwhelmingly influenced by the performance of a single dominant risk factor – interest rate risk.

Bag the antiquated Agg.

Considering the 2021-22 regime change, it’s prudent to loosen the Agg’s shackles and embrace actively managed fixed income funds that invest in multiple bond sectors, maturities and effective durations. Given the macroeconomic environment and changes in the debt markets, the Agg has outlived its usefulness.

Managing the risks of restrictive monetary policy alongside mixed fundamentals and a potential recession underscores the case for multi-sector, multi-factor fixed income strategies. In this environment, it’s limiting to confine the core of a portfolio to Agg-based strategies. It’s also risky, given the uncertain prospects for higher rates and how duration effects drove negative returns in 2022. Active strategies can tailor a portfolio’s duration and sector allocations in a more flexible, market-aware fashion. Active duration management, sector allocations, and security selection can better defend against rate and credit risks than indexed core Agg bond funds.

Rick Roche, CAIA, is managing director of Little Harbor Advisors, LLC, a position he has held since April 2013. He is a 42-year veteran of the industry, dually registered adviser, and CAIA charter holder. Over the course of four decades, he’s trained thousands of credentialed financial advisors in practice management and a variety of timely investment topics (85+ CE programs over the last 4-5 years). His bio is here.

The information contained in this publication was obtained from sources believed to be reliable, but its accuracy cannot be guaranteed. Opinions expressed herein are those of the author, Rick Roche, and not those of Little Harbor Advisors (LHA).

¹ Wikipedia, “Pig in a poke”, accessed on 7/14/23.

² Bloomberg, “The most widely used fixed income indices globally”, accessed on 07/19/23.

³ Vlastelica, R., “The cracked benchmark? Why some investors want a new standard for bonds”, MarketWatch, July 24, 2017.

⁴ Duration is a function of yield (coupon) and maturity. Think of it as weighted average time to get principal repaid. When rates, – yield or coupon – go up, duration goes down; when maturities are up, duration is up. Source for the Agg’s historical average “effective duration” is FEG Insight, “2023 Fixed Income Market Outlook” by Keith Berlin.

⁵ Rauterberg, G., & Verstein, A., “Index Theory: The Law, Promise and Failure of Financial Indices”, University of Michigan Law School Scholarship Repository, 2013.

⁶ “Bloomberg Fixed Income Index Methodology”, Bloomberg Index Services Limited (BISL), Aug 24, 2021.

⁷ Phillipps, J., “A brief history of indices”, CityWire, June 4, 2020.

⁸ Ibid., Rauterberg, G., & Verstein, A., page 26.

⁹ Index Industry Association (IIA), “Highlights of Index History Timeline”, July 6, 2022.

¹⁰ Redding, Rick, CEO, Index Industry Association, IIA survey results from June 2018, personal correspondence rcv’d on 07/21/2023.

¹¹Cambrian explosion: About half a billion years ago, life on Earth changed dramatically within a relatively short period of time. A rapid increase in the diversity of biological species took place in Earth’s ancient oceans which is when most major animal groups started to appear in the fossil record.

¹² Bailey, T., “There are now 70 times more stock market indices than listed stocks in the world”, Interactive Investor, Jan 24, 2018.

¹³ Ibid., Rauterberg, G., & Verstein, A., page 11.

¹⁴ Ibid., Dick-Nielsen, J., and Rossi, M, page 21.

¹⁵ Ibid., Rauterberg, G., & Verstein, A., page 4; and Vaughan, L., “BBA Officials Worried That Libor Manager Was `Out of His Depth’”, Bloomberg, Oct 15, 2015.

¹⁶ Webb, K., Cantor Fitzgerald, “Understanding Benchmarks Concepts”, Dec 31,2016.

¹⁷ Dick-Nielsen, J., and Rossi, M., “The Cost of Immediacy for Corporate Bonds”, The Society for Financial Studies, July 24, 2018.

¹⁸ Ibid., Dick-Nielsen, J., and Rossi, M.

¹⁹ J. P. Morgan, “Build Stronger Fixed Income Portfolios”, Winter/Spring 2023.

²⁰ Ibid., “Bloomberg Fixed Income Index Methodology”, Aug 24, 2021.

²¹ Fane, E. & Williamson, R., Morningstar “The Return of the Bond Market”, Mar 16, 2023.

²² U.S. Bureau of Labor Statistics, “Consumer Price Index: 2022 in Review”, Jan 17, 2023.

²³ Bloomberg Terminal, Agg composition data accessed on 7/12/2023.

²⁴ Cremers, Martijn, Choi, J., & Riley, T., “Why Have Actively Managed Bond Funds Remained Popular?”, June 2023.

A message from Advisor Perspectives and VettaFi: Advisors who don’t grasp the implications of artificial intelligence (AI) risk being left behind. Join VettaFi for an AI Symposium on August 30th at 11 am ET and hear from thought leaders and experts about how AI will transform the way investing works. Register here!

Read more articles by Rick Roche