Diversification in direct lending

Introduction

In portfolio construction, diversification is paramount owing to its capacity to mitigate risk, particularly in the face of tail events. This is especially true in direct lending, where investors tend to focus on downside protection: Too much concentration in a single borrower can amplify return volatility and cause material negative impacts on the overall portfolio’s performance. A comparable risk exists when a portfolio’s loans are acquired through a single general partner (GP), essentially tethering the portfolio’s outcomes to the performance of that sole GP.

This article employs our proprietary data to highlight the merits of diversification through two dimensions: the number of positions and the number of GPs within a portfolio. We also attempt to assess these advantages quantitatively by utilizing measures such as internal rate of return (IRR) and loss rate distributions. Our findings indicate that a more diversified approach corresponds with less severe tail events.

Data

Our analysis relies on our proprietary database of middle-market loans, which encompasses more than 22,500 loan tranches and represents approximately $678 billion in effective drawn amount. To ensure the robustness of the study, we include only direct US first-lien loans. This approach helps us eliminate region-specific characteristics, minimizes discrepancies arising from different capital structures, and ensures that GPs pursuing other strategies (e.g., opportunistic or distressed lending) are excluded.

Direct loans often feature multiple tranches, each ranking pari passu. A single transaction could therefore encompass a term loan, a revolving facility and a delayed draw term facility. To nullify any potential bias stemming from such structures, we have exclusively incorporated the “main tranche” of each transaction. This refers to the tranche with the largest drawn amount, which is most often a term loan. In addition, we removed GPs that made fewer than 10 loans in our dataset to avoid any potential bias in our analysis. We also excluded loans with incomplete data.

This data cleaning process left us with a final dataset made up of 5,030 loans, representing more than $185 billion in effective drawn amount. These loans were made between 2006 and 2021 and were sourced from a diverse pool of 41 GPs (Figure 1).

Methodology

The initial segment of our analysis centers on evaluating the influence of borrower concentration on a portfolio’s IRR (net of losses and gross of GP fees and costs) and loss rates. Our aim is to determine if reducing single borrower concentration within a portfolio reduces the severity of tail risks and to quantify the potential impacts.

To do this, we implement a methodology that is applied throughout subsequent sections by using either our whole final database or a specific subset of it.

Initially we create a portfolio by randomly selecting a prespecified number of loans from our pool of assets, with replacement (e.g., a portfolio of 100 randomly selected loans).¹ Upon building this portfolio, we estimate its IRR and loss rates by using the simple average of its positions. Finally, after storing the relevant metrics, we build a new portfolio and repeat the process. The number of iterations must be high enough so that we can properly determine the distribution of the outcomes (i.e., IRR and loss rates) for each set of portfolios containing a predefined number of loans. After various trials, we concluded that 100,000 portfolios suffice to build a stable distribution. The choice of opting for the simple average implies that equal weight is assigned to all loans within each portfolio.

¹ The use of replacement means that selected loans remain accessible within the asset pool and can potentially be chosen multiple times. It is a common practice and ensures that each loan is always being selected from the same distribution.

Explore more insights on private debt

Podcasts

RPM—Ep.36 | Building a private debt portfolio with Christian Frei

Market insights

Private debt: key observations 1H23

Whitepapers

A systematic approach to private debt allocation in institutional portfolios