Chapter 12: CFA Level 1 Financial Statement Analysis
Building robust financial models for valuation and forecasting
Financial modeling is the process of building a forecast of a company's future financial performance. The primary goal is typically to perform a valuation. The process starts with forecasting revenue and then flows through the rest of the financial statements.
It's essential for analysts to be aware of their own biases and external factors like competition, economic trends, and technology, as these can significantly impact the model's accuracy.
Key Purpose: Financial models serve as the foundation for investment decisions, providing a structured approach to value companies and assess future performance under different scenarios.
Constructing a robust model involves a systematic process, starting with a broad industry view and progressively narrowing down to company-specific details.
Analyze the industry structure using Porter's Five Forces:
Identify the company's different business segments, their respective revenue streams, and their contributions to overall profit.
This is the core of the financial model, where you project the company's future income.
| Component | Forecasting Approach |
|---|---|
| Revenue Forecast | Combine a top-down approach (considering broad economic factors like regional GDP growth) with a bottom-up approach (analyzing company-specific segment trends). Key drivers to consider include changes in volume, price, product mix, and foreign exchange rates. |
| Cost of Goods Sold (COGS) | Often forecasted as a percentage of revenue (COGS margin). Consider factors like pricing power and supply constraints, which can affect the gross margin. For example, high prices due to limited supply might improve the gross margin, even if COGS is declining. |
| Selling, General, & Administrative (SG&A) | Also typically forecasted as a percentage of revenue. This may increase if the company is expanding its distribution network or increasing advertising spending. |
| Profitability Metrics |
From the above forecasts, you can calculate:
|
| Non-Operating Items & Taxes | Include any non-operating income or expenses. Forecast the corporate income tax by considering historical effective tax rates and any expected changes. |
Analyst Awareness: Analysts must be vigilant about their own psychological biases, which can systematically skew forecasts. Recognizing these biases is the first step to mitigating them.
Description: Overestimating one's own forecasting abilities and the precision of the forecast.
Mitigation: Rigorously track and analyze past forecast errors. Use scenario and sensitivity analysis to consider a range of outcomes.
Description: Believing one can control or influence outcomes that are, in fact, uncontrollable. This often leads to overly complex models.
Mitigation: Acknowledge limitations, focus on the most relevant data, and avoid information overload.
Description: Failing to fully incorporate new information, instead clinging to prior beliefs or forecasts (anchoring).
Mitigation: Consciously evaluate the impact of new information without being too heavily anchored on past data.
Description: Classifying new information based on past experiences or familiar patterns, leading to misinterpretations.
Mitigation: Consider all relevant information, including underlying base rates and overall context.
Description: Seeking out information that confirms existing beliefs while ignoring contradictory evidence.
Mitigation: Actively seek out opposing viewpoints and consult independent sources for unbiased assessment.
Changes in the general price level have a significant but varied impact on company forecasts.
Companies with strong pricing power can better protect their profit margins by passing cost increases to customers through higher prices.
Falling input costs can boost margins, but companies must balance price cuts to maintain volume without eroding profitability.
Key Insight: Inflation modeling requires distinguishing between nominal and real growth. Always assess whether revenue growth is driven by volume or price increases, and whether cost increases can be passed through.
The explicit forecast period (e.g., 3, 5, or 10 years) should be long enough to allow the company to reach a "normalized" level of earnings. The choice depends on:
Since a company is assumed to operate indefinitely, terminal value captures cash flows beyond the forecast horizon. Common methods:
Small changes in assumptions have large effects:
Example 1:
FCF = $100M, discount rate = 10%, growth rate = 3%
Terminal Value = $1.43 billion
Example 2:
Growth rate increases to 4%
Terminal Value = $1.67 billion (17% increase)
Always perform sensitivity analysis on terminal value assumptions.
Financial statement modeling is important for valuation. Master these areas:
SGR Formula Confusion: SGR = ROE ×Retention Ratio, NOT payout ratio. If payout ratio is 40%, then retention ratio is 60% (1 - 0.40). Don't confuse the two.
Perpetual Growth Rate Selection: The perpetual growth rate (g) should be conservative, typically 2-4% tied to long-term GDP growth. Using a company's historical growth rate (e.g., 8-10%) is unrealistic and implies permanent growth above the economy.
g r in Gordon Model: If growth rate equals or exceeds the discount rate, the terminal value formula produces negative or infinite values. The perpetual growth rate must be less than the discount rate for the model to work.
Actual Growth Exceeding SGR: If a company's actual growth exceeds its sustainable growth rate, it must either issue new equity (diluting shareholders), increase leverage (higher debt/equity), improve ROE (higher profitability/efficiency), or reduce dividends (increase retention).
Terminal Value Sensitivity: Small changes in the perpetual growth rate have large impacts on terminal value. A 1% increase in g can increase terminal value by 10-20% or more. Always perform sensitivity analysis on terminal value assumptions.
Behavioral Biases: Overconfidence bias = overestimating forecast accuracy. Illusion of control = believing you can influence uncontrollable outcomes. Confirmation bias = seeking only confirming evidence. Conservatism bias = slow updating of forecasts. Representativeness bias = classifying based on stereotypes.
Forecasting Technique Selection: Sensitivity analysis changes ONE variable at a time. Scenario analysis changes MULTIPLE variables simultaneously (recession/base/boom). Monte Carlo simulation runs thousands of scenarios using probability distributions.