DSCR in Project Finance

Introduction

In this article we introduce the DSCR calculation; in particular it's application in a project finance financial model. We intend you to read this with the accompanying Excel workbook so you can review the layout of the calculations and the actual formula - rather than just read about it. If you find this useful and you want to understand more about the intention, interpretation and market nuances of the DSCR and how it is used in debt-sizing ask us about our two day, 100% hands-on, case study based, Project Finance Analysis course, delivered in person by Nick Crawley.

Download Workbook Download PDF Training Courses

The Debt Service Cover Ratio (DSCR) is one of the two primary ratios that are heavily relied upon for the structuring and monitoring of a project finance transaction. The reliance of project financing on the financial modelling of the DSCR makes it an essential area of expertise for a project finance modeller. The DSCR is a Covenant; these are parameters and conditions that must be met for the smooth running of the project finance asset. Covenants are defined in the Term Sheet or Facility Agreement.

For project finance financial modelling it is crucial to thoroughly grasp the concept and application of the DSCR. Whilst the ratio itself is simple, when structuring transactions and analyzing scenarios it becomes so prominent that as, especially as an analyst, the DSCR is a dimension you need to be able to think in.

After 20 years+ of model auditing experience we can say that the DSCR is a common source of errors in project finance financial models and one of the first things we would check. To help you manage this situation this post also introduces you to what those common errors are as well as what are the success factors for a smooth and confident DSCR modelling process!

The DSCR formula

The DSCR is on the face of it, a simple ratio. It is cash available to service debt divided by the debt service (principal, interest and usually financing fees) in corresponding periods. The DSCR is calculated in the operational phase of the project finance model; usually on a quarterly basis but that depends on the timing of the covenants and the nature of the project financing.

The DSCR, in an operational period i is calculated as:

Where:

CFADS = Cashflow Available for Debt Service
P = Principal due for repayment
I = Interest and fees due for cash payment (not accrued)

A DSCR of 1.00x means that the cash available only just meets debt service. 1.50x means there is 50% more the bare minimum. A healthy DSCR is generally ~1.75x to 2.50x but this depends on the nature of the debt facility and the specifics of the project financing.

However, it is rarely this simple and instead often straddles time periods before, including and often ahead of the period i. So the more general formula is

Where:
i = the period in which you are measuring the DSCR
t1 = the beginning of the measurement window
t2 = the end of the measurement window

t1 is usually < i and t2 is commonly equal to i or maybe i+2 in the case of a look-forward measurement. This approach is coded in workbook so take a look.

The intention

The incorporation of the DSCR in a project finance financial model is to identify where the weaknesses are in the projects ability to meet it’s debt service obligations, period by period over the life of the debt facility. The DSCR profile gives the projects owners, advisors and lenders the ability to tailor the debt parameters to ensure the coverage profile meets their credit and risk objectives – for a lender it is usually an exercise in "maximizing the minimum" of the DSCR time series. A more advanced treatment might maximise the average DSCR whilst constraining the minimum DSCR to be above a hurdle value and the debt still being repaid within a given timeframe. The actual approach is heavily dependent on the often unique circumstances of the project as well as the sophistication of the structuring team.

There are two important floors in relation to the DSCR covenant; which if breached have commercial consequences. It is helpful to show both of these on any plot of the DSCR to visually convey headroom.

DSCR Lock-up – typical values are 1.10 to 1.20x
DSCR Default – typical values are 1.05 to 1.10x

Put simply, when the DSCR in any period falls below lock-up the shareholders are not able to distribute cash dividends; the intended consequence being that any free cash that would have been paid out in dividends is accumulated in the bank account to provide support to the in-period cashflows and for the coming debt service. As you can imagine this is a challenging trade-off between debt and equity parties.

How is the DSCR calculated?

There are several variations which cover periods of time, rather than only in a single period. The most common variations are

4 quarter rolling look-back. The most common and generally fariest approach
2 quarter look-back and 2 quarter look-forward

The second approach is a common source of circular references but is a useful approach for structuring and monitoring projects with volatile cashflows.

The commercial rationale of variations

Variations which take different periods into account, both backwards and forward looking, are tailored to respond differently in different circumstances. For example, in the case of a negative shock to CFADS potentially due to an unexpected one-off cost a 4-quarter lookback DSCR will smooth out the impact over 4 quarters whereas an in-period DSCR would be the most sensitive. Lenders and borrowers often do not want covenants to be unduly sensitive as it causes often unhealthy nervousness on a quarter to quarter basis when in reality both parties to the loan know that these things do happen in the normal course of business and they only want to be at emergency restructuring meetings when it is absolutely critical.

How to structure the DSCR calculation

In the workbook you will see the extracts required for the DSCR calculation; being a CFADS line and elementary debt facility - both which feed the DSCR calculations. Whilst the inputs can be changed this workbook is all about the DSCR rather than modelling debt facilities; so these are intentionally simplistic.

Screenshot-DSCR-Extracts-(1).JPG

The next step is to set-up the DSCR calculations, starting with bringing all required lines into one local area to avoid long referencing.

We systemise the DSCR calculation to be between dynamic periods. This approach has two benefits, firstly it facilitates setting up a new variation in a low-risk manner which helps experiment whilst structuring; secondly there is only one formula rather three different ones to check / understand.

In the accompanying workbook you will see our formula is:

=IFERROR ( CFADS / Total Debt Service , 0 ) * FLAG

The three important points to take-away here are:

The use of the FLAG. A binary flag (i.e: 1 or 0) which nulls the result if the debt has been repaid.
The simplicity of the ratio has been preserved it is A/B with some conditions. We only use brackets when needed.
We use IFERROR to avoid a Div/0 result beyond the life of the debt. Cleaner than IF < Tenor.

Minimum and average DSCR

There three essential outputs from the DSCR time series, these are

The average DSCR
The minimum DSCR
The date the minimum DSCR occurs

Investigate the formula in the workbook and you will see

Average DSCR is the arithmetic average of the time series not the ratio of CFADS / P+I over the life the loan - this is not the average and a common mistake. We use AVERAGEIF() to exclude the zeros after the end of the debt facility.

The Minimum DSCR is calculated using MINIFS(), again to exclude zero values beyond the life of the loan. You can do this with an array too but MINIFS is less daunting for most users.

The Date of the minimum DSCR is calculated using INDEX(Date range,MATCH(Min DSCR, DSCR time series,0)). If you are sculpting to a constant DSCR this aspect sometimes needs to be double checked or even ignored depending on the situation.

These calculations lead up to the all important visual presentation of the ratios, CFADS, debt service and the headroom (in DSCR space) over the DSCR Lock-up and DSCR Default covenants.

Analysing this plot we see that a shock in CFADS is treated in three different ways because of the windows (t1,t2) of each of the ratios. Firstly, note that the plot is clearly titled, labelled, has a clear legend and can print in black and white and still be read. One can argue that a line plot is not a mathematically correct way to represent the discrete DSCR time series; experience says it is is the most easy to cognitively digest and is generally the market standard. Let's take a look at how the three DSCR calculations vary with a sudden, single period, drop in CFADS.

The DSCR in-period ratio (fine dots) drops the quickest and recovers the quickest. It gave no foresight and dropped to the lowest level of all three. Very sensitive.

The DSCR 4-quarter look-back (dashes) gives no warning, drops significantly less in the period of the lower CFADS and recovers more slowly

The DSCR 2-quarter look-back and 2-quarter look-forward, gives advance warning of problem (if the issue could be forecast), doesn't drop as low as the in-period DSCR and recovers.

Summary

This is an elementary insight into the DSCR, it is a deceptively challenging subject especially when combined with LLCR, debt sculpting and return constraints. If you found this helpful and would like to learn about other aspects of Project Finance Modelling or Advanced Financial Modelling then you would love our training courses! Check them out here or just give us a call.

I hope that was useful – smooth and happy modelling!