Financial modeling is the process of creating a mathematical representation of a company's or investment's financial performance to analyze its current state, forecast future outcomes, and aid in decision-making.

Historical data may not accurately predict future performance due to changing market conditions, economic factors, and company-specific developments. However, certain fundamental metrics.

While numerous macroeconomic factors can influence financial analysis, the focus of this post is to establish a robust method for identifying and excluding definitively poor investments, aiming for 100% precision in flagging "bad" stocks, even if it doesn't capture all potentially good ones. This conservative approach is adopted to minimize the risk of false positives, ensuring that investors avoid stocks that are highly likely to underperform, even if it means potentially missing out on some good opportunities.

Stocks selection: what are GOOD and BAD companies?

In this analysis, we're working with specific definitions of "good" and "bad" stocks:

Bad stocks = those companies that either went bankrupt or experienced a severe negative return exceeding 90%. It's important to note that this classification only applies to companies with available financial data, ensuring our analysis is based on concrete information.

While identifying failed companies is straightforward, recognizing truly good businesses is more nuanced. Therefore, I've manually selected companies from a well-established and widely accepted list of superior businesses to represent the 'good' category in this analysis.

In particular, I've used this list of 227 bad companies and this list of 31 good companies for the current study.

Goal of this empirical study

Despite the imbalance between good and bad stocks, our focus on achieving 100% precision rather than accuracy aligns well with a conservative investment strategy, as it prioritizes avoiding false positives (mistakenly identifying bad stocks as good) over potentially missing some good opportunities.

What about financials manipulations?

While it's true that companies can potentially manipulate their financial statements, our study's methodology inherently accounts for this risk. By directly fetching and analyzing the reported financials, we incorporate any potential manipulations into our assessment. This approach is robust because it captures the actual data presented to investors and regulators.

It's crucial to emphasize that this distinction is rooted in the businesses' financial fundamentals and operational performance, not in their stock market prices or short-term market fluctuations. This approach allows us to focus on the intrinsic qualities of the businesses rather than potentially volatile market sentiments.

Procedure

The study uses a simple algorithm to analyze financial metrics of stocks labeled 'good' or 'bad'.

It systematically tests combinations of these metrics and their value ranges, aiming to identify criteria that select only good stocks.

The algorithm prioritizes solutions using fewer metrics while maximizing the number of unique stocks identified.

This approach seeks to create a simple yet effective screening method for stocks based on financial data.

Results

The analysis yielded a set of financial criteria that consistently identified good stocks while excluding all bad ones.

Note: The following figures are presented as ratios rather than percentages. So for example, a value of 0.1 in the alert criteria should be interpreted as 10%.

These criteria are:

1. Net profit margin is between 0.1 and 0.88
2. Earnings growth rate is between 0 and 0.2
3. Gross margin is between -0.9 and 1
4. ROIC to ROE ratio is between -0.9 and 0.7
5. Debt-to-asset ratio to ROA is below 100
6. Dilution is below 0.1
7. Asset turnover is below 1

This combination of metrics successfully identified 19 unique stocks out of the 31 "good" stocks in our dataset (61.3%).

In terms of individual data points, where each point represents a single year of data for a stock, the criteria captured 119 "good" data points out of a total of 1057 (11.3%).

Crucially, this screening method achieved 100% precision, excluding all "bad" stocks and their associated data points from the selection. This means our screening method not only identified a majority of the good stocks and captured multiple years of strong performance for these companies, but also completely avoided false positives. These criteria provide a conservative screening method that, based on historical data, effectively filtered out all known bad companies while retaining a significant subset of consistently good performers, demonstrating its robustness in identifying quality investments.

How to get notified when those criteria are met?

If you'd like to receive a weekly notification when the above conditions are met, please click here to set up a weekly alert.

---

Discover Valu8.app

Valu8.app is an NLP-powered stock screener that lets you input complex criteria in plain English.

It runs weekly algorithmic scans and emails you matching stocks, replacing manual screening and constant market monitoring.

Try Valu8 today and save time on your investment research!

Learn more...