Financial models and their nuances (Part 1)

Financial models are tools used by analysts to estimate the future financial performance of any financial asset – be it equities, bonds, or complex derivatives and securities. Just like in fashion industry where a model is used to show the look of a fashion garment and value it, in finance, they are used to see how the cash flows of the asset will look like and then value them. In essence a financial model shows the analyst the future cash flows/profits etc of any financial asset.

A financial model shows the analyst the future cash flows/profits etc of any financial asset

While there can be many types of financial models for different asset classes, the scope of this post will be to convey how we think about financial models for publicly listed companies. Such models are built by fundamental analysts to estimate the future financial performance of the company over a longer time period (typically 3 years in future, although some more detail oriented analysts can extend these models to even 10 years). These models help the analysts to value the company using absolute valuation methodologies such as DCF, or use relative valuations on some parameters such as EPS, EBITDA, Book value etc.

At DistrictD we believe that these expectations of the future performance of a company play a very important role in its value and how that is expected to perform. And hence our core product revolves around these models. To us they are the bread-and-butter for all seasoned investors.

Typically a financial model for a company is made up of 4 segments

  • Historical data
  • The Business model
  • Assumptions of drivers
  • Valuations

In this post we will cover the first two of these aspects.

The Historical data refers to past data of financial performance, the operating drivers/parameters of the company and any global/macro variables and indicators. While past performance is never a guarantee of the future, it does provide lots of insights about the business. A few of the insights can be of the following nature

  • Base size of operations and assets of the company that it has attained so far
  • Trends on recent performance of operating and financial drivers
  • Long term cyclical trends of the business
  • Draw correlations on business performance based on historical events (both macro and internal).
  • Compare historical performance of similar companies, to derive conclusions on strength of business and management

Above all, Historical data serves as a base for projecting the future performance of the company.

Next comes the Business model of the company we are building the model for. The business model captures how a company functions financially – i.e. how it derives revenues, what are its costs and how it is financed. In effect it is a set of mathematical calculations that we use to estimate how various business drivers interact with each other to estimate the company’s profitability. Say for example, modelling for an IT outsourcing company would mean the set of calculations that multiply the onshore and offshore headcounts with the respective billing rates to arrive at the total revenues and then consider the respective costs of the two cost centres appropriately to get the margins of the business. So as a user if you want to see how much of a margin benefit comes through increased offshoring or say INR depreciation, the business model helps estimate it.

An important aspect of embedding the business model in the financial model lies in zeroing down on the parameters/drivers that would be relevant for the company. This involves understanding each company’s business in certain details – such as where do the revenues come from, how asset intensive is it to get those revenues, what costs are incurred to get them, regulatory costs, taxes, capital structure etc.

We at DistrictD place a very large emphasis on getting the operational model of the business in place, which is getting the revenue and operating costs in as granular details as possible. This is because we believe structural investors always look at steady state operational performance outside of any one-time or exceptional impacts. Having said that, our models do allow users to test for all the other aspects as well – including capital structure (by increasing/reducing leverage, share issuances/buybacks and dividends), capital expenditure, working capital, taxes etc.

The quality of a good financial model lies in three things. Firstly it is in identifying the key drivers for the business model, ones which are tracked by management, investors and other stakeholders. Building a financial model based on parameters that are not significant to the current paradigm of the company, means it is impractical to use it. For example order book is a very important driver for a lot of infrastructure companies whereas unit sales is a more relevant driver for Automotive OEM companies. Secondly it is in getting the right interactions between these drivers in the business model. If the drivers are incorrectly linked the model will throw incorrect results, and may actually result in the analyst trying to make incorrect driver inputs to get what he thinks is a correct expectation. Thirdly the drivers should be independent of one-another to as much extent as possible. We will discuss a bit more about independence of drivers in the next post.

Finally as a closing to the Business model discussion for now, I would like to end with some of the limitations that we work with.

  • There are alternate ways to model the same company. For example a telecom business can be modelled in terms of “Subscribers and Average Revenue per User (ARPU)” or as ”Data and Voice revenues”, both of which are exclusive to each other. This means it is hard to create a single model that incorporates both these methods.
  • We work with public information disclosed by the company. This means that many times we do not have accurate information on certain business drivers to be able to model the business in a better fashion.
  • We try to construct models that are based on the drivers that the market follows or ones that management discloses. This means we may not have data for all the drivers required, and may have to make certain assumptions to complete the entire picture.
  • The drivers that we work with might not be independent of one another. This implies that if a driver changes it has an impact on some of the other input drivers as well. “Mutual independence of drivers” is a detailed topic and deserves a blog post of its own.
  • A lot of profitability decisions are driven by market forces and management discretion, both of which cannot be modelled as driver.
  • It is very difficult to model the impact of macroeconomic conditions
  • Companies keep evolving over time and so does their business model


Leave a Reply