Syndicated Revolver Part 1.5
Welcome to Part 1.5 of my Applied Machine Learning in Finance blog series. Feel free to skip to Part Two if you’re here for the models. Last week, we introduced the syndicated revolver and the problem. This week, we’re going to import some data and build some predictive models.
Before we dive into importing, I wanted to elaborate on the Syndicated Revolver. I’ve received a lot of questions about the details of the process and I hope this will answer them.
Last week as an example we used Uber as a company, but this week we will be using Coca-Cola Bottling Co. Consolidated, Coke, instead.
What’s the downside?
A syndicated revolver is very costly for the bank. Public companies the size of Coke have to comply with many layers of regulations to ensure that they are stable and reliable entities. They are also only able to enter “committed” revolvers which declare that the bank that collaborates with Coke needs to have the funds earmarked for the duration of the revolver.
Coke is looking for $1 billion over five years to expand into Mars. They think they have the numbers to make it work, but the laws force them to take out a loan. They also want a card up their sleeve in case their projected numbers don’t pan out.
This is outrageously costly for the bank for a few reasons. The bank has to pay their treasury in order to get access to $1 Billion. The treasury will charge them healthy fee for this privilege. Deposit fees must be paid because you can’t just park $1 billion in your savings account. Maintaining and supervising that amount of money costs a lot of money. There are also administrative duties that further increase the costs.
Another issue for the bank is that a company like Coke probably will never need to draw down on that loan unless they find themselves in a worst-case scenario. The thorough layers of regulation that companies must grow to that size creates a lot of safety nets. The bank might be on the hook for the billion dollars over the five years. They will pay a lot of money to have access to the money, more money to make sure that the money is being maintained properly and at the end the company doesn’t even use the money.
How do the banks finance this?
Banks have different groups which use different financial instruments to cover the loans that they are offering companies.
On the more exotic side a group might use luxury cars or rare art as collateral and on the more traditional side you can raise money from the street by issuing debt new equity or taking deposits however all of those things cost money so banks have to be very careful in the deals that they pursue.
Why do banks partner up?
How do you eat an elephant? Banks partner to share burden of maintaining the money. If a bank can split a $1 Billion deal in 100 pieces, it isn’t as scary a task.
What is the goal of the model?
In our model we will be looking to strike the biggest deal while inviting the least amount of participants.
Think of an airline company that wants to overbook a flight but not to the point of being a trending topic on Twitter. There is a fragile balance that must be monitored closely. Our goal is to send out a few invites as possible while still creating enough capital to fund the loan.
Further down the line we will add a sentiment analysis to try to target companies that will be spending a lot of money and are more likely to drive down on the debt, but first we’ll start with raw numbers.
