Liability-driven investment policies and asset management decisions are those largely determined by the sum of current and future liabilities attached to the investor, be it a household or an institution. As it purports to associate constantly both sides of the balance sheet in the investment process, it has been called a “holistic” investment methodology.


Traditionally, liability-driven investment policies (LDIs) have been reserved for institutional investors, like pension plans, with clearly defined future cash-flow obligations.  However, who is to say that LDI is not appropriate for retail investors as well?  After all, most people have fixed, monthly expenses such as car loans, mortgages, and student loan debt.  Beyond the monthly obligations, most retail investors have a degree of certainty around expected future cash outlays such as housing down payments or college tuition for children.  Why shouldn’t an LDI approach be taken in these cases?

LDI approaches generally use a mixture of fixed-income securities (or derivatives), which have defined future cash flows, seeking to match future liabilities.  Unfortunately, defining an appropriate LDI benchmark is an art in-and-of itself, with the need to closely match the interest rate risk, credit risk, and yield curve risk of future obligations.  Too much tracking error to “liability returns” means that future cash flows are not adequately hedged.

Fortunately, Barclays and Russell teamed up to deliver the Barclays-Russell LDI Index Series (data), a suite of LDI strategies that seek to simultaneously reduce tracking error versus “liability returns” while retaining their status as investable solutions.  In other words, no obscure, non-replicable OTC derivatives.  The long-short of the situation is that there is a series of corporate-bond based indices designed to hedge interest rate sensitivity at different levels.

So how do we access these indices as a retail investor?  Given that all the marketing materials for this suite is still directed towards pension plans, it is not surprising that they haven’t been packaged as ETFs or mutual funds and it is unlikely that we’ll have the capacity to license and replicate the index with individual securities.

Fortunately, the investability mandate of the suite, combined with the proliferation of fixed-income ETFs, means that we might be able to replicate the indices.  In other words, we might be able to create a suite of ETF-based portfolios that closely track the returns of the LDI indices.

A quick peak under the hood into the index construction methodology document tells us that the indices pull from a universe of credit bonds and (in the case of the 16-year LDI index) US Treasury STRIPS.  Googling quickly turns up 4 ETFs that should meet our needs: the Vanguard Short, Intermediate, and Long-Term Credit Bond ETFS (VCSH, VCIT & VCLT) and the Vanguard Extended Duration ETF (EDV).  These ETFs, together, should match the universe while giving us the flexibility to create mixtures to match durations.

To actually replicate the index, we’ll run an optimization process that seeks to minimize the standard deviation of the differences between our replicating portfolio’s returns and the index’s returns – also known as tracking error.  Specifically, at the end of every month, we’ll create an exponentially-weighted covariance matrix from the prior 12 months and minimize the following tracking-error formula:


subject to the constraint that all weights are between 0-100% and sum up to 100%.  For a more detailed treatise on where this formula comes from, click here.

We can see the results of this process below, where the blue line represents the actual index and the orange line represents the replicating portfolio.

16-year ldi

14-year ldi

12-year ldi

10-year ldi

8-year ldi

6-year ldi

Clearly, the replication isn’t perfect: tracking errors average approximately 1% annualized.  But as a naive and raw approach, it isn’t half bad.  We could likely improve our tracking error by utilizing more than 12 data-points in our covariance matrix construction, implementing a shrinkage estimator, and even more explicitly taking into account the actual index construction methodology (e.g. trying to match rebalance periods, making our optimization subject to duration constraints and ensuring that the STRIPs only appear in the 16-year solution).

An important consideration for the feasibility of this methodology is how stable our allocations are.  We can see in an example, below, that for the 16-year LDI replication portfolio, our turnover is reasonable, clocking in at less than 50% annualized.  Certainly not the most passive strategy, but not the most active either.


As retail investors continue to pursue innovative means of managing their investments, they may do well to take a lesson from institutional investors: matching your future liabilities is a great foundation.  And by combining a simple, quantitative technique with available, liquid ETFs, we can leverage the intellectual horsepower of institutional-calibre solutions that have been packaged as indices.

Corey is co-founder and Chief Investment Officer of Newfound Research, a quantitative asset manager offering a suite of separately managed accounts and mutual funds. At Newfound, Corey is responsible for portfolio management, investment research, strategy development, and communication of the firm's views to clients. Prior to offering asset management services, Newfound licensed research from the quantitative investment models developed by Corey. At peak, this research helped steer the tactical allocation decisions for upwards of $10bn. Corey holds a Master of Science in Computational Finance from Carnegie Mellon University and a Bachelor of Science in Computer Science, cum laude, from Cornell University. You can connect with Corey on LinkedIn or Twitter.