Income Statement Forecast

Overview


Time Series Analysis


The basic framework for forecasting each entry in an income statement is to think of income statement as a time series and apply the corresponding methods. (Technically, the income statement becomes a vector valued time series).

Typically, analysts will use simplifying assumptions. As an example, when forecasting revenue, it is generally assumed that the revenues grow at an average growth rate (plus or minus some noise). 


let sdata = st.sort(data, p=>p.date);
let growthRate = $list(sdata).map((p,i,data)=>{
                          if(i>0) return p.revenue/(data[i-1].revenue)
                       })
                      .filter(p=>p!== undefined)
                      .average();
				
let forecastedRevenue = sdata[sdata.length-1].revenue * growthRate;
Try it!


Calculated Values


Some values such as Gross Profit Margin are simply defined by a formula of other income statement items. These values do not need to be forecast (assuming the underlying variables are being forecast)

Pre-defined Values


Depreciation occurs on a pre-defined schedule. That is, for all assets held at the beginning of the accounting cycle, this number will be known in advance (unless management changes its accounting treatment).

If the analyst knows the depreciation schedule of the firms assets, she will not necessarily have to forecast this item. However, she may still apply forecasting techniques as a shortcut.

Budgeted Values


Some of the entries on the income statement are not entirely random, that is, they are chosen. For example, the research and development expense line item is (largely) chosen by management. For analysts working within the firm, the research budget can be specified in advance and not modeled. For analysts outside the firm, if management does not make its plans known, these items will have to be forecast in the same way as the other items. Oftentimes, they are assumed to be constant.

Dependence Structure


One of the challenging aspects of forecasting the income statement is that the items on the statement are not independent. For example, the cost of goods sold will have a strong correlation with revenue. Correlations can be modeled using standard time series tools, but often they are just assumed to have some fixed relationship with another item on the income statement. For example, the cost of goods sold is often assumed to be a fixed percentage of sales revenue.

The average proportion of cogs/sales can be computed in a similar fashion to the revenue above. 


let sdata = st.sort(data, p=>p.date);
let growthRate = $list(sdata).map(p=>return p['cost of goods sold']/p.revenue)
                      .average();
				
let forecastedCogs = forecastedRevenue * growthRate;


Some items will have a dependence on items that are on the balance sheet. As an example, interest expense is computed as an interest rate multiplied by a debt balance. As an approximation, the debt balance can be taken off the balance sheet of the prior period, however, if new borrowings occur (or if debt is paid down) during the current accounting period, this will only be an approximation.

A more accurate estimate can be attained by creating a circular dependence in the forecasted values of the financial model and finding the limiting value. (see interest expense for more information)

Hidden Variables


Often times, there are variables not included on the income statement that are easier to forecast and which imply the items on the income statement. As an example, instead of forecasting sales revenue, a firm could forecast the number of items of product sold, and the price of each item. These variables may be easier to forecast if the firm has a robust marketing analytics department which forecasts them. (see demand forecasting and pricing)

Forecast Items


  • Sales - typically assumed to grow at some growth rate, which is estimated as an average of prior periods. If the firm has a sophisticated marketing department, sales may come from a forecast produced using demand forecasting and price analytics. Additionally, if the firm is engaging in a marketing campaign, purchase response analysis may provide some additional insight.

    Sales may also be impacted by problems with the firms supply chain including inventory bottlenecks.
  • Cost of Goods Sold - usually taken to be some fixed percentage of sales. This may be impacted by current trends in inflation. (see us inflation for current data)
  • Operating Expenses :
    • Research and Development - chosen by management. Often, assumed to be constant.
    • Selling General and Admin - usually taken to be a fixed percentage of sales. A more accurate estimate of this item can be created if a firm has an accurate cost allocation methodology. In particular, separating fixed and variable costs and modeling the cost curve (ala experience curve) can yield insights.
    • Depreciation - often taken to be a fixed percentage of total assets. sometimes taken to a constant
  • Other Income - usually taken as a fixed amount equal to average of prior values
  • Other Expenses :
  • Extraordinary Items - usually taken as a fixed amount equal to average of prior values, unless a particularly large one time item has ocurred recently
  • Taxes -

Sample Models


Contents