home / skills / louloulin / claude-agent-sdk / graham-value-investing

graham-value-investing skill

safe

This skill applies Graham's deep value framework to identify safe, discounted investments through Net-Net screening and margin-of-safety based valuation.

npx playbooks add skill louloulin/claude-agent-sdk --skill graham-value-investing

Review the files below or copy the command above to add this skill to your agents.

Files (6)

SKILL.md

5.9 KB

---
id: graham-value-investing
name: Graham深度价值投资
description: Ben Graham的深度价值投资方法,强调安全边际和深度折价购买
version: 1.0.0
author: InvestIntel AI Team
dependencies:
  - financial-data
  - valuation-framework
tags:
  - value-investing
  - fundamental-analysis
  - ben-graham
  - margin-of-safety
---

# Graham深度价值投资

**作者**: InvestIntel AI Team
**版本**: 1.0.0
**最后更新**: 2026-01-11
**标签**: value-investing, fundamental-analysis, ben-graham, margin-of-safety
**依赖**: [financial-data, valuation-framework]

---

## 📖 技能描述

Ben Graham的深度价值投资方法，强调安全边际和深度折价购买。本技能实现Graham的核心估值方法和筛选标准。

### 核心理念

1. **安全边际** (Margin of Safety) - 以低于内在价值30-40%的价格购买
2. **Net-Net筛选** - 最严格的深度价值标准
3. **Graham公式** - 简化的内在价值计算
4. **防御型投资** - 优先保护本金，其次追求收益

---

## 🎯 投资标准

### 1. 安全边际要求

```yaml
minimum_margin_of_safety: 0.30  # 30%最低安全边际
preferred_margin_of_safety: 0.40  # 40%首选安全边际
```

**计算公式**:
```
安全边际 = (内在价值 - 当前价格) / 内在价值
```

### 2. Graham公式估值

**基础公式**:
```
V = EPS × (8.5 + 2g)
```

其中:
- V = 内在价值
- EPS = 每股收益
- g = 预期增长率 (以小数表示，如0.05表示5%)
- 8.5 = 零增长公司的合理倍数

**修正公式** (当利率显著偏离历史平均时):
```
V = EPS × (8.5 + 2g) × (4.4 / Y)
```

其中:
- Y = 当前AAA级公司债券收益率
- 4.4 = 历史平均AAA债券收益率

### 3. Net-Net筛选标准

最严格的Graham筛选标准，适用于深度价值投资:

**Net-Net营运资本定义**:
```
Net-Net = 流动资产 - 流动负债 - 总负债
       = 现金 + 应收账款 + 存货 - 总负债
```

**买入标准**:
```yaml
net_net_price_ratio: 0.666  # 价格 ≤ 2/3 Net-Net价值
```

**筛选条件**:
- 当前价格 ≤ 2/3 × Net-Net每股价值
- 流动资产 > 总负债
- 无重大诉讼或监管风险
- 过去10年持续盈利

### 4. 财务健康度要求

```yaml
# 流动性指标
current_ratio: 2.0  # 流动比率 ≥ 2.0
quick_ratio: 1.0    # 速动比率 ≥ 1.0

# 负债水平
debt_to_equity: 0.5  # 债务权益比 ≤ 0.5
current_assets_to_total_liabilities: 2.0  # 流动资产/总负债 ≥ 2.0

# 盈利稳定性
earnings_consistency: 10  # 过去10年持续盈利
positive_earnings_years: 9  # 过去10年中至少9年盈利
```

---

## 🔍 投资流程

### 步骤1: Net-Net筛选

1. 计算Net-Net营运资本
2. 计算Net-Net每股价值 = Net-Net / 总股本
3. 筛选: 价格 ≤ 2/3 Net-Net价值

### 步骤2: 财务健康检查

1. 流动性检查: 流动比率 ≥ 2.0
2. 负债检查: 流动资产 ≥ 2 × 总负债
3. 盈利检查: 过去10年至少9年盈利

### 步骤3: Graham公式估值

1. 获取过去3-5年平均EPS
2. 估算合理增长率g (保守估计)
3. 计算内在价值 V = EPS × (8.5 + 2g)
4. 验证安全边际 ≥ 30%

### 步骤4: 综合评分

```yaml
评分标准 (总分100):

1. 安全边际 (30分)
   - 50%+: 30分
   - 40-50%: 25分
   - 30-40%: 20分
   - <30%: 0分

2. 财务健康 (25分)
   - 流动比率 > 2.5: 10分
   - 2.0-2.5: 8分
   - 1.5-2.0: 5分

   - 债务权益比 < 0.3: 10分
   - 0.3-0.5: 8分
   - 0.5-0.7: 5分

   - 盈利连续性 (10年全盈利): 5分

3. 盈利质量 (25分)
   - ROE > 20%: 15分
   - 15-20%: 12分
   - 10-15%: 8分

   - 自由现金流 > 净利润: 10分
   - 80-100%: 8分

4. 估值吸引力 (20分)
   - P/E < 10: 20分
   - 10-12: 15分
   - 12-15: 10分

   - P/B < 1.0: 10分
   - 1.0-1.5: 5分
```

**投资决策**:
- 80分以上: 强烈买入
- 70-79分: 买入
- 60-69分: 观察
- 60分以下: 不投资

---

## 📊 估值示例

### 示例1: Net-Net机会

**公司数据**:
- 股价: $10
- 流动资产: $100M
- 总负债: $30M
- 总股本: 10M

**计算**:
```
Net-Net = $100M - $30M = $70M
Net-Net每股 = $70M / 10M = $7
买入价上限 = 2/3 × $7 = $4.67
```

**结论**: 股价$10 > $4.67，不满足Net-Net标准 ❌

### 示例2: Graham公式

**公司数据**:
- 股价: $50
- EPS (过去3年平均): $5
- 预期增长率: 5%

**计算**:
```
V = $5 × (8.5 + 2×5%) = $5 × 8.6 = $43
安全边际 = ($43 - $50) / $43 = -16.3%
```

**结论**: 安全边际为负，不投资 ❌

### 示例3: 满足标准

**公司数据**:
- 股价: $25
- EPS (平均): $4
- 预期增长率: 6%
- 流动比率: 2.8
- 债务权益比: 0.2
- ROE: 18%

**计算**:
```
V = $4 × (8.5 + 2×6%) = $4 × 8.62 = $34.48
安全边际 = ($34.48 - $25) / $34.48 = 27.5%
```

**评分**:
- 安全边际: 20分 (27.5% < 30%)
- 财务健康: 25分 (优秀)
- 盈利质量: 23分 (ROE 18%)
- 估值吸引力: 10分 (P/E = 6.25)

**总分**: 78分 → 买入 ✅

---

## ⚠️ 风险提示

### Graham方法的局限性

1. **忽略增长质量**: 只看增长率数量，不看质量
2. **过时标准**: Net-Net标准在现代市场中极其罕见
3. **行业差异**: 未考虑不同行业的合理估值倍数差异
4. **账面价值失真**: 现代公司中无形资产占比大，账面价值可能失真

### 适用场景

✅ **适用于**:
- 传统制造业
- 周期性行业底部
- 小型被忽视公司
- 熊市恐慌期

❌ **不适用于**:
- 高科技成长股
- 生物医药未盈利公司
- SaaS/订阅制业务
- 互联网平台

---

## 🔗 相关资源

### 经典著作
- 《证券分析》 (Security Analysis) - Benjamin Graham & David Dodd
- 《聪明的投资者》 (The Intelligent Investor) - Benjamin Graham

### 相关技能
- [Buffett质量价值投资](../buffett-quality-value/SKILL.md) - Graham方法进化版
- [Kelly仓位管理](../kelly-position-sizing/SKILL.md) - 仓位优化
- [Munger思维模型](../munger-mental-models/SKILL.md) - 多元思维模型

---

## 📝 变更历史

### v1.0.0 (2026-01-11)
- 初始版本
- 实现Graham公式估值
- 实现Net-Net筛选
- 实现综合评分系统

Overview

This skill implements Benjamin Graham’s deep value investing framework focused on margin of safety and buying at deep discounts to intrinsic value. It codifies Net-Net screening, the Graham valuation formula (with bond yield adjustment), and a multi-factor scoring system to support buy/observe/sell decisions. The skill is designed for disciplined, capital-protecting value investors seeking systematic screening and valuation guidance.

How this skill works

The skill computes intrinsic value using Graham’s formula (V = EPS × (8.5 + 2g), adjusted by prevailing AAA bond yield when needed) and measures margin of safety against the current market price. It performs a strict Net-Net calculation (current assets minus total liabilities, per share) and applies liquidity, leverage, profitability, and stability filters. Finally it aggregates weighted sub-scores (safety, financial health, earnings quality, valuation) into a 0–100 score with thresholds for buy/hold/avoid.

When to use it

Screening undervalued, traditional or cyclical companies where balance-sheet metrics matter
Seeking disciplined entry points with defined margin-of-safety thresholds (≥30%)
Evaluating small or neglected stocks with strong asset coverage or recurring earnings
Running systematic value checks during bear markets or panic-driven sell-offs
Filtering out growth-oriented or intangible-heavy businesses unsuitable for balance-sheet value methods

Best practices

Use conservatively estimated growth rates (g) based on 3–5 year EPS averages and industry context
Verify net-net calculations on a per-share basis and confirm no off-balance-sheet liabilities or major contingencies
Require liquidity and leverage thresholds (current ratio ≥2.0, current assets ≥2× total liabilities) before considering purchase
Combine quantitative scores with qualitative checks (management integrity, litigation, regulatory risk)
Treat Net-Net hits as high-conviction but rare opportunities; expect low frequency in modern markets

Example use cases

Identify small-cap manufacturing companies trading below 2/3 of Net-Net per-share value for potential deep-value purchases
Run batch valuation across a watchlist to highlight stocks with ≥30% margin of safety per Graham formula
Score candidates for a concentrated, capital-preserving value portfolio using the 0–100 rubric
Reject tech or high-intangibles firms automatically and focus analyst time on balance-sheet-rich targets
Monitor bond yield adjustments to Graham valuation when AAA yields deviate materially from historical averages

FAQ

Does the method work for tech or SaaS companies?

No. The framework relies on tangible assets and earnings history; firms with large intangible value or no consistent earnings are not suitable.

How conservative should the growth rate g be?

Use a conservative estimate based on historical EPS trends and industry outlook; default to a low single-digit rate unless justified.