What is “Modeling Assumptions”?  

Modeling assumptions form one of the four components in the forecasting process within the realm of financial modeling – which is used to project or forecast a company’s financial performance over a specific period of time (eg. 5 years, 10 years, etc.). These assumptions play a pivotal or central role in forecasting financial data, as the forecasts are based on these assumptions. For example, if historical sales have gone up 10% every historical year, the assumption can be for this to continue (at 10%) into the forecasting period. Therefore, based on assumptions, if year 0 (historical period) sales were US$100, this year’s forecast sales would be US$110. 

The four components of the forecasting process stepwise are: 

  1. Inputting publicly available or private historical data 
  2. Constructing ratios and statistics predicated on this historical data  
  3. Creating assumptions for future performance based on the historical performance 
  4. Forecasting financial data based on assumptions

The forecasting process necessitates the preparation of the company’s financial statements – balance sheet, income, and cash flow statements, based on modeling assumptions.

Key Learning Points  

  • Modeling assumptions refer to the creation of assumptions about a company’s future growth based on analyzing its historical performance
  • The set of assumptions made about future business conditions drive the forecasts of a company’s financial statements and is the basis of building a financial model
  • The modeler has design choices to make while constructing a financial model and the forward assumptions which drive the model output

Future Assumptions – Salient Features

The key point to note regarding modeling assumptions is that a financial model is basically built on this set of assumptions. The assumptions will decide the expected future business conditions that drive the forecasts of a company’s balance sheet, cash flows, and income statement (for example, forecasts of line items such as revenues and earnings). 

Further, a very important aspect of modeling assumptions, is that items such as earnings and cash flows – must be forecast properly into the future, as the outlook for financial performance of a company largely determines the intrinsic value of its stock.

If we see the historical sales going up by 5% in every historical period, we may assume that the future sales in the forecast period will also go up by 5%. If last year’s sales were 100, step forward sales would be 105. Assumptions may be for sales to grow faster or slower than the previous year.

The other place to look for future assumptions is within industry research in case there are any factors which may have a positive (or negative) impact on this company. If we were using a 5% assumption for the future sales growth we may choose to increase it to reflect robust economic conditions.  

The modeler has several design choices to make when constructing a model – how many excel sheets to use, where to put the model assumptions (i.e. treatment of the forward assumptions which drive the model output), and how to format negative numbers. 

The first design choice is what to do with our assumptions. The options are to either put all the forward assumptions in one place in the model i.e. they are ‘blocked’ together. In other words, forward assumptions regarding revenue growth assumption, operating costs & revenues assumption and taxes as % of profit before tax assumption are all together. Alternatively, some modelers prefer to have the assumptions spread throughout the output or model. The former option is more reader-friendly. 

Blocked assumptions are particularly important in large models – particularly having all in one tab. For example, if you have the revenue growth assumption (say 3%), then we have the output revenues together – so you have the driver (3%) and the output together. The benefit here is that the output and the driver being visible together means that we can actually see what is driving the output. 

Modeling assumptions should always be explicitly documented in a model so that a user can see them easily and then change them when required. It is best modeling practice to use an assumption cell to build the forecast data. This means that the output can be changed simply by amending the content of the assumption cell (for example sales growth changing from 4% to 5%). It is never appropriate to embed the assumption in the output formula. Doing this means that the driver is invisible and cannot be changed without adjusting formulas.  

Most modelers will position the forward assumptions besides the equivalent historical ratio so that the trend line can easily be sense checked. Any variation in this trend should be adequately explained with a well-documented comment. 

Ratios and Forward Assumptions – Generating Assumptions From Historical Data

Here, we see two years of actual historical data and 5 years of estimated (forward) assumptions – Year 1 – Year 5. We can see that in the historical period for Year 1 and Year 0, annual sales growth was 6.2% and 6.6% respectively. Looking at this historical data helps us to understand the business and where it is likely going. 

We can use the historical figures to come up with the forward assumptions as well as using additional research for this purpose such as information from the company’s management or our own opinion. The revenue growth assumptions are laid out below. These forward assumptions will drive the forecast model for five (or 10) years into the future.

Ratios and Forward Assumptions

Forecast Model – Income Statement and Balance Sheet 

Given below is an example, where the income statement and balance sheet for the forecast period are built, with historical data and some ratios and statistics are calculated. This includes revenue growth, costs as % of revenues, inventories as % of costs and long-term debt change. These in turn help in coming up with the forecast assumptions for the same from periods 1 to 3. 

Therefore, for the forecast process, we first have the historical data, then ratios computed using the historical data, thereafter we have the forecast assumptions and ultimately our financial forecasts.  

Having stated this, there is something missing before we start and that is the historic net income (income statement) has to be computed  (i.e. revenues – costs). There are several steps to complete the balance sheet:  

  • The total assets have to be calculated too (i.e. cash + inventories) 
  • The sum of total liabilities and total liabilities plus equity of the historical period has to be computed 
  • A balancing check is required to ensure that total liabilities + equity – total assets = 0

Next, to calculate the revenues for the forecast period, we will multiply the last (historical) year’s revenue with the revenue growth assumption (5%) and so on to get forecast revenues. For costs, we see the cost as % of revenues assumption of 85% (i.e. 85% of US$110.3 (revenue) and so on to get the forecast costs. We can then calculate the net income for the forecast period. 

Moving down to the balance sheet, the cash assumption is already given. For calculation of inventories for the forecast period, we use the inventories as a % of costs (13%) assumption and multiply it by the costs in the forecast year (period 1 to 3). Thereafter, adding these up, we get the total assets for the projected years.

Lastly, we want the total assets to be equal to total liabilities and equities. To arrive at total liabilities, we compute long-term debt for the forecast period, using the long-term debt change assumption (0%) i.e. no change from last year in the first projected year. There is no specific assumption for equity. So, we take the net income from the income statement and add it to last year’s equity (i.e. US$16.5 + US$14 = US$30.5)  and so on. We see that the balance sheet balances. 

This is efficiently how assumption modeling is used to derive future forecasts in financial modeling.

Forecast Model

Forecast Model

Forecast Model