How to Build a Three-Statement Financial Model

A three-statement financial model is a single connected spreadsheet that projects the profit and loss, the balance sheet and the cash flow statement so that they move together as one system. You build it in a fixed order: the P&L first from revenue and operating assumptions, then the balance sheet items such as working capital, fixed assets and debt, and finally the cash flow statement, which is not typed in by hand but falls out of the period-to-period changes in the other two.

The point of the exercise is not the output. It is the structure. Once the three statements are correctly linked, the model becomes a machine for testing assumptions: change one input, and every downstream number updates in a consistent, arithmetically honest way. This guide walks through the build order, the linkages that turn three tables into one model, and the two headaches that trip up almost everyone the first time.

Why three statements, and why linked

Each statement answers a different question. The P&L asks whether the business made a profit over a period. The balance sheet asks what the business owns and owes at a point in time. The cash flow statement asks where the cash actually went, which is often a very different story from the accounting profit.

A company under Ind AS can report a healthy net profit and still run short of cash, because profit recognises revenue when it is earned, not when it is collected. Working capital, capital expenditure and debt repayment all consume cash without touching the profit line in the same period. A model that only forecasts the P&L cannot see this. A linked three-statement model can, because the cash flow statement reconstructs cash from the changes in every other line.

The discipline that holds the whole thing together is the accounting identity: assets equal liabilities plus equity, in every single period. If your projected balance sheet does not balance, you do not have a rounding problem. You have a broken link, and the model is telling you so.

The build order

Work top to bottom, statement by statement. Resist the urge to jump around. The order below exists because each step depends on the outputs of the one before it.

Step	Statement	What you build	Key inputs
1	P&L	Revenue, margins, EBITDA, depreciation, EBIT	Volume and price assumptions, cost ratios
2	Balance sheet	Working capital, fixed assets, capex, debt, equity	Days-based ratios, capex plan, financing plan
3	P&L (finish)	Interest expense, tax, net profit	Debt schedule, effective tax rate
4	Cash flow	Operating, investing, financing cash flows	Changes between periods in steps 1-2
5	Balance sheet (close)	Cash balance, retained earnings	Closing cash from step 4, net profit from step 3

Notice that the P&L and balance sheet each appear twice. That is deliberate. You build the operating part of the P&L first, but you cannot finish it until the debt schedule tells you the interest expense. You lay out the balance sheet structure early, but the cash line and retained earnings can only close once the cash flow statement has run. This back-and-forth is normal and is where the linkages live.

Building the P&L

Start with revenue, because it drives almost everything else. Pick a build that matches how the business actually earns money. A generic goods business might model revenue as units sold multiplied by an average realisation. A services business might model it as billable capacity multiplied by a utilisation rate. Keep the driver visible as a separate assumption cell, never buried inside a formula.

From revenue, work down through the cost structure. Model the major cost lines as ratios or as their own drivers: raw material as a percentage of sales, employee cost as a growth rate, and so on. That gives you EBITDA. Subtract depreciation and amortisation to reach EBIT.

Then stop. You cannot compute interest yet, because interest depends on the debt balance, which lives on the balance sheet you have not built. Leave interest, profit before tax and net profit as placeholders and move on. Coming back to close the P&L after the balance sheet is a feature of the method, not a mistake in it.

Building the balance sheet

The balance sheet is where most of the real modelling judgement sits. Break it into a few blocks.

Working capital. Model receivables, inventory and payables using days-based ratios tied to the P&L. Receivables might be set as a number of days of revenue, inventory as days of cost of goods, payables as days of purchases. When revenue grows, these balances grow, and that growth quietly ties up cash. This is the single biggest reason profit and cash diverge.

Fixed assets and capex. Roll the asset base forward with a simple schedule: opening gross block, plus the year’s capital expenditure, gives closing gross block. Accumulated depreciation grows by the depreciation charge you already put in the P&L. The net block feeds the balance sheet, and the depreciation figure ties directly back to the P&L. This shared depreciation number is one of the model’s core links.

Debt and equity. Lay out a debt schedule: opening balance, drawdowns, repayments, closing balance. Equity grows by retained earnings, which is last year’s balance plus this year’s net profit less any dividend. Share capital usually stays flat unless you model a fresh issue.

At this stage the balance sheet will not balance, and that is expected. The plug that makes it balance is cash, and cash comes from the statement you build next.

Letting the cash flow statement fall out

Here is the part that feels like magic the first time it works. You do not forecast the cash flow statement directly. You derive it, line by line, from the changes you have already built.

Operating cash flow starts from net profit, adds back depreciation because it is a non-cash charge, and adjusts for the change in working capital. Investing cash flow is mostly capex. Financing cash flow is debt drawdowns and repayments, less dividends. Sum the three, add opening cash, and you get closing cash.

That closing cash figure drops into the balance sheet as the cash line. And now the accounting identity does its job: if every link is correct, assets equal liabilities plus equity automatically. You do not force it. If it balances, your logic is sound. If it does not, a change somewhere is not flowing through consistently, and the imbalance points you at the error.

The essential linkages, stated plainly:

• Net profit flows to retained earnings on the balance sheet, connecting the P&L to equity.
• Depreciation ties the P&L to fixed assets, appearing as a charge on one and reducing the asset base on the other, then added back in operating cash flow.
• Working capital changes drive operating cash, which is why a growing, profitable business can still be cash-hungry.
• Debt drives interest expense, closing the loop back into the P&L.

The two headaches everyone hits

The balance sheet will not balance. This is the most common failure, and it is almost always a missing or double-counted link, not a formula typo. Work backwards. Check that net profit flows fully into retained earnings, that depreciation is consistent in both places it appears, that every balance sheet movement has a matching line in the cash flow statement, and that closing cash from the cash flow actually feeds the cash line. One of those will be broken.

Interest creates a circular reference. Interest expense depends on the average debt balance. Debt depends on whether the business generated or burned cash. Cash generated depends on net profit, which depends on interest. The formula chases its own tail. There are two standard fixes. You can enable iterative calculation so the spreadsheet resolves the loop, though this can hide errors and should be used with care. Or you can break the circle by charging interest on the opening debt balance rather than the average, accepting a small approximation in exchange for a clean, transparent calculation. For a learning model, the opening-balance approach is usually the safer choice.

A practical takeaway

A three-statement model is a reasoning tool, not a crystal ball. Its output is only ever as good as the assumptions you feed it, and its real value is that it forces those assumptions to be consistent with each other across all three statements at once. Build it in order, keep every driver in its own visible cell, and treat the balance sheet balancing as your continuous test that the logic holds. Once the structure is sound, spend your time where it matters: pressure-testing the handful of inputs that actually move the outcome, one at a time, and watching how the whole system responds.