PitchBook blog

Your resource for all things PitchBook
Limited Partners

Best practices for optimizing cash flow forecasting

Curious how to forecast a cash flow for private market funds? Here we explore cash flow forecasting models for optimizing predictions.

When David Swensen left Wall Street to manage Yale’s endowment, he arrived with a novel idea—replacing the formulaic model of stocks and bonds the endowment had traditionally employed in favor of private market funds, like PE and real estate. And decades later, this sacrificed liquidity has reaped major returns for the university—in his 36 years, Swensen grew the Yale endowment from $1.3 billion to over $40 billion, averaging a 13.7% compound annual gain.

Other university endowments quickly followed in Swensen’s footsteps, incorporating Yale’s model of private market allocations into their own portfolios. Institutional investors of all types were soon turning their curiosities toward the private market. Today, pension funds, fund managers, foundations, and sovereign wealth funds all look to the private markets in search of high returns and greater diversification, with many allocating 15%, 20%, even 30% or more of their portfolio to private market funds.

But the private markets are not all big returns and paydays for limited partners (LPs). The attributes that characterize the high returns of these investments are also the reasons behind their unique risks and challenges. The illiquid nature of private market funds coupled with the variability of capital calls and distributions poses a challenging obstacle for LPs—balancing meeting capital calls with maximizing returns becomes a difficult needle to thread.

While maintaining adequate cash on hand is necessary for avoiding capital call defaults, an equally important consideration for the LP is to avoid allocating unnecessary capital to low-return, liquid assets like treasury bills, or even worse—a cash account.

Enter cash flow forecasting.

Forecasting cash flows for private market funds is the best way institutional investors can mitigate this fundamental issue. An effective cash flow forecasting model enables LPs to better predict the timing and amount of capital calls and distributions, reducing their exposure to inferior returns. In this blog, we look at the challenges and importance of cash flow forecasting for LPs and touch on our cash flow forecasting model for private market funds developed by PitchBook analysts and detailed in our Allocator Solutions: Cash flow forecasting and commitment pacing report.

What is cash flow forecasting for private market funds?

Unlike traditional investments where capital is put to work immediately, private market funds are typically closed-end, in which LPs make an upfront commitment to a fund that is drawn over time at a variable rate.

During the fund’s investment period (typically the first 5 years), general partners (GPs) make investments by calling committed capital from LPs. As the fund matures, drawdowns tend to slow, as fewer new investments are made.

Later in the fund’s life, GPs begin to exit their investments, distributing the proceeds from these exits to limited partners. This pattern is observed in the standard J curve graph, which portrays the generalized net cash flow of capital in private market funds.

Although this pattern is observed in most private market funds, there is a high degree of variability around this curve with differing schedules of contributions and distributions that are affected by a variety of factors.

Fund size, available exit opportunities, and dry powder are just a few of the elements that can affect the timing and size of contributions and distributions. For example, larger funds with more dry powder are more likely to deploy larger capital calls over a longer period.

Fund strategy also plays a significant role. Private debt funds typically see the fastest drawdowns of any private capital strategy, with more than 60% of their total commitment being called on average in the first two years compared to just 37% for PE.

Cash flow forecasting models attempt to address these nuances in private market funds that are not accurately represented on a simplistic J-curve. Through employing historical data, statistical analysis, and probability simulations, LPs can leverage cash flow forecasting templates to better predict the amount of capital that will be called or distributed over their investment’s lifespan.

Why is cash flow forecasting important?

Effective cash flow forecasting models provide LPs with a heightened understanding of cash flow patterns, enabling them to better plan for both capital calls and distributions while enhancing returns by keeping a smaller allocation of their uncalled commitments in low-yield, liquid assets. In the best scenarios, these models can help LPs recycle distributions for upcoming contributions.

These benefits make forecasting cash flows a vital task for LPs, as even reducing the amount of capital in treasury bills or cash by a low margin can have a significant impact on returns. And with a good model, LPs can successfully do this without worrying about missing a capital call, which has serious repercussions such as interest penalties, reputational damage, and reduced future distributions.

It is important to note that even the most sophisticated LPs who implement highly accurate cash flow models will inevitably have to set aside some portion of uncalled capital to a low-yielding asset, but a good cash flow model can effectively reduce this exposure.

What should a cash flow forecast include?

Cash flow forecasting models for private market funds typically calculate three outputs:

  • Contributions, or capital calls an LP will have to make over the fund’s life
  • Distributions, or returns an LP will receive over the fund’s life
  • Net asset value (NAV) of the portfolio, which describes the total value of the underlying assets held in the portfolio

Each of these values are modeled with a formula that provides a forecast for both their timing and value. Time is usually calculated by quarter and size as a percentage. For example, a model for contributions would output the percent of committed capital expected in a given quarter.

The best cash flow forecasting models provide users with the flexibility to incorporate inputs specific to the fund they are trying to model, such as the fund’s length, investment period, vintage year, return assumptions, cumulative cash flow statistics, and others. Additionally, more sophisticated models will take an empirical rather than theoretical approach, drawing on historical data rather than estimations for calculations.


PitchBook's guide to overcoming four allocator workflow challenges

Learn how LPs can use data-driven insights to uplevel their investment strategies—with trusted research, benchmarks, and portfolio allocation tools.
Download the guide

How many methods of cash forecasting are there?

When it comes to cash flow modeling, there are two main approaches: theoretical and empirical. Here we’ll look at the Takahashi Alexander Model, the most common theoretical model used in the industry alongside PitchBook’s empirical cash flow forecasting model.

What is the Takahashi Alexander model?

During his time managing the Yale Endowment, David Swensen worked closely with Dean Takahashi and Seth Alexander to develop a cash flow forecasting model for illiquid alternative assets. This model, commonly referred to as the Takahashi Alexander Model has become the most popular model LPs use for cash flow modeling.

The Takahashi Alexander model uses a variety of inputs to project capital contributions, distributions and the NAV for a given quarter (t). These outputs are referred to as D(t), NAV(t) and C(t) and the inputs used to calculate these values include:

  • Rate of contribution
  • Total capital commitment
  • Life of the fund
  • Bow factor, describing changes in the rate of distribution over time
  • Annual growth rate
  • Yield percentage
  • Paid-in capital, or the amount of capital an LP has already paid into the fund

The Takahashi Alexander model is theoretical in nature, meaning it is based on hypothetical rather than empirical or historical data, and the inputs such as the bow factor, growth rate and rate of contribution, are based on assumptions or estimations rather than observed numbers in a robust historical data set.

Additionally, the Takahashi Alexander model is deterministic, meaning for every input you put into the model, you receive a singular output, which will often be quite different from the real-life values. Following cash flow forecasting best practices, it is preferred to supplement forecasts with a probabilistic component, most commonly a Monte Carlo-style simulation, to provide perspective on the potential variability of the forecast.

PitchBook’s cash flow forecasting model for private market funds

Our models for contributions and distributions are empirically based, developed from the extensive cash flow and historical fund data available in the PitchBook Platform.

Our cash flow forecasting models for distributions and contributions utilize a baseline, “normalized” cash flow profile for each private market asset class—private equity, venture capital, private debt, etc.—that is calculated by extracting historical data on each fund strategy from the PitchBook Platform.

LPs can then leverage the normalized cash flow profiles for a given asset class to model the expected contributions, distributions, and NAVs of the funds they are committed to through introducing specific inputs into our models for contributions and distributions.

Some of the inputs that are used in our models include:

  • Fund strategy
  • Total committed capital
  • Fund length
  • Investment period
  • TVPI, or the total value of paid-in capital
  • A given quarter

Our model is flexible, allowing LPs to introduce supplementary inputs to model a fund mid-life, such as its current age and capital called and distributions to date. Additionally, since GPs provide LPs with regular updates regarding the expectations for a fund, our models allow these predictions to be implemented into our formulas to generate customized cash flow profiles.

Since our models produce a singular output as described for the Takahashi Alexander Model, we supplement them with Monte Carlo-style simulations, providing LPs with a range of potential cash flows and allowing them to approximate the worst-case scenarios for capital call events.


Allocator Solutions: Cash flow forecasting and commitment pacing

Take an in-depth look at our cash flow forecasting models along with detailed cash flow forecast examples and probability simulations of private market funds.

Download report

What are the limitations of cash flow forecasting?

While effective models can help LPs improve cash flow forecasting for their funds, varying models have certain limitations that are important to note. Specifically, the main limitations of our cash flow forecasting model include:

  • Since our normalized cash flow profiles are aggregated across vintage years within each strategy, our models assume that historical data will represent future cash flows
  • Our forecast estimates are based on quarterly figures and for simplicity, assume that the cash flows occur at the end of a quarter—to calculate forecasts on more specific timelines, additional data manipulation would be required

Despite these constraints, our model improves upon the Takahashi Alexander model as well as other off-the-shelf models available today, providing LPs with a method for cash flow forecasting based off a robust historical data set. Ultimately, LPs can leverage our model to better predict the cash flows of their funds and minimize the amount of capital they have to allocate to low-return investments.


Ready to optimize your cash flow forecasting?
Explore PitchBook's portfolio forecasting solutions