Financial modelling and integrated statements

What a model is actually for

A financial model is not a prediction. It is a decision machine. Its job is to translate a set of assumptions into the consequences for profit, cash, and the balance sheet, so a manager can see which assumption the decision really depends on.

That reframing changes how a model is built. A good model makes its assumptions visible and easy to change, keeps the accounting identities intact, and reports the few outputs that drive the decision. A bad model hides assumptions inside hard-coded numbers and produces a beautiful sheet that no one can defend.

Build from drivers, not guesses

The fastest way to a defensible forecast is to express every line in terms of a driver. Revenue is units multiplied by price, or last year's revenue grown at a stated rate. Cost of goods sold is a percentage of revenue. Payroll is headcount multiplied by average cost. Working capital is a number of days.

Once the model is built from drivers, a scenario is not a second copy of the spreadsheet; it is a different set of driver values. Growth of 12% versus 6%, a margin of 40% versus 36%, a collection period of 45 days versus 60. The structure stays fixed and only the inputs move.

The three statements must stay linked

The statements are one system, not three reports. Net income flows into retained earnings on the balance sheet. Depreciation reduces profit but not cash, so it is added back in the cash flow statement and accumulated as a contra-asset on the balance sheet. Capex leaves cash and becomes a fixed asset. Growth absorbs working capital, which is why a profitable company can still run out of cash.

A model stays honest through a single discipline: the balance check. Assets minus liabilities minus equity must be zero in every period. Equality is not a formatting nicety; it is the proof that every flow has been routed to the right place. When the check breaks, the error is almost always a flow that was recorded but never carried through to the balance.

Circularity, and how to live with it

There is one genuinely awkward loop in every integrated model. Interest expense depends on how much debt is outstanding. Debt depends on how much cash the business generated. Cash generation depends on net income, which depends on interest expense. The circle is real, and it is why spreadsheets offer iterative calculation.

Most practical models break the loop deliberately. The simplest and most defensible approach is to charge interest on the opening debt balance, so interest for the year is known before the cash flow is computed. The model is one period slower and completely auditable. Save iterative calculation for cases where the loop is material, and always show a clear error if it fails to converge.

A worked year

Nova Manufacturing forecasts revenue of ৳100,000,000 and expects it to grow 12%, to ৳112,000,000. Gross margin is 40%, so gross profit is ৳44,800,000. Operating expenses are 25% of revenue, ৳28,000,000, leaving EBIT of ৳16,800,000.

The company holds ৳40,000,000 of debt at 9%, so interest is ৳3,600,000 and profit before tax is ৳13,200,000. At a 25% tax rate the charge is ৳3,300,000, leaving net income of ৳9,900,000. Paying out 40% as dividends leaves ৳5,940,000 in retained earnings.

On the cash side, add back depreciation of ৳6,000,000, subtract a ৳2,000,000 increase in working capital and ৳8,000,000 of capex. Free cash flow is 9,900,000 + 6,000,000 − 2,000,000 − 8,000,000 = ৳5,900,000. If the model cannot explain why that number is not equal to net income, the integration is not finished.