What Is the Harmonic Mean?
The harmonic mean is a type of average calculated as the reciprocal of the arithmetic mean of the reciprocals of a set of numbers. In simpler terms: take the reciprocal of each value, compute the arithmetic mean of those reciprocals, and then take the reciprocal of that result. The harmonic mean is appropriate for averaging rates and ratios — situations where the numerator is fixed and the denominator varies — such as speed (miles per hour), price-earnings ratios, or cost per unit. It is always the smallest of the three Pythagorean means (harmonic ≤ geometric ≤ arithmetic), and it gives less weight to large outliers and more weight to small values than does the arithmetic mean. While less commonly used than the arithmetic or geometric means, the harmonic mean is essential in specific financial applications, including the calculation of certain stock indices, portfolio performance measures, and cost averaging problems.
How the Harmonic Mean Works
The formula for the harmonic mean (H) of n numbers is: H = n / (1/x₁ + 1/x₂ + ... + 1/xâ‚™). A classic illustration: a car travels from City A to City B at 60 mph and returns from City B to City A at 40 mph. What is the average speed? The arithmetic mean of 50 mph is incorrect — the car spent more time traveling at 40 mph because the return trip took longer. The harmonic mean gives the correct answer: 2 / (1/60 + 1/40) = 2 / (0.0167 + 0.025) = 2 / 0.0417 = 48 mph. More time was spent at the slower speed, pulling the true average below the arithmetic mean. The harmonic mean correctly captures this temporal weighting. For financial data, the harmonic mean is useful when averaging ratios where the denominator is the variable of interest. For example, to find the average P/E ratio of a portfolio, the harmonic mean — not the arithmetic mean — gives the correct P/E of the combined earnings stream. If a portfolio holds equal dollar amounts of two stocks with P/Es of 10 and 30, the correct portfolio P/E is the harmonic mean: 2 / (1/10 + 1/30) = 15, not the arithmetic mean of 20.
Real-World Example: Dollar-Cost Averaging
An investor practices dollar-cost averaging, investing $1,000 each month into a mutual fund. In month 1, the fund's share price is $50, buying 20 shares. In month 2, the price drops to $25, buying 40 shares. The arithmetic mean price is ($50 + $25) / 2 = $37.50. But the investor bought more shares at the lower price — 40 shares at $25 versus 20 shares at $50. The average cost per share is total dollars invested divided by total shares purchased: $2,000 / 60 shares = $33.33. This is the harmonic mean of the prices: 2 / (1/50 + 1/25) = 2 / (0.02 + 0.04) = 2 / 0.06 = $33.33. The harmonic mean correctly reflects that the investor's average purchase price is closer to the lower price because more shares were accumulated at that price. This example illustrates both the harmonic mean's application to cost averaging and the broader principle that dollar-cost averaging naturally leads to a lower average cost per share than the arithmetic average of prices — a key benefit of systematic investing.
When to Use the Harmonic Mean
Use the harmonic mean when you need to average rates (miles per hour, cost per unit, items per dollar, pages per minute), financial multiples (P/E ratios, EV/EBITDA multiples across a portfolio), or any ratio where the numerator is constant and the denominator varies. The key diagnostic question: does the quantity being averaged have a fixed numerator? If the trip distance is fixed (same journey each way), use the harmonic mean for average speed. If the dollar amount invested in each stock is fixed (equal-weighted portfolio), use the harmonic mean for the portfolio P/E. If neither numerator nor denominator is fixed, the geometric mean is often appropriate. In practice, the harmonic mean appears most frequently in finance in the calculation of the equal-weighted harmonic mean P/E ratio for indices and portfolios, in the measurement of dollar-cost averaging cost basis, and in certain performance attribution calculations. It is a specialized tool that, when correctly applied, gives the right answer where the arithmetic mean would give the wrong one.
Why the Harmonic Mean Matters
The harmonic mean is not a mathematical curiosity — it is the correct solution to a specific and recurring class of averaging problems. Using the arithmetic mean where the harmonic mean is required produces systematically biased (typically overstated) results. In portfolio analysis, using arithmetic instead of harmonic means for P/E ratios can make a portfolio appear cheaper than it actually is. In performance measurement, using arithmetic means for rates can overstate actual performance. The three Pythagorean means — arithmetic, geometric, harmonic — each answer a different question about the data, and knowing which question you are asking determines which mean to use. The harmonic mean answers: what is the average rate or ratio, properly accounting for the varying weight of each observation? In a financial world awash with ratios and rates, that question arises more often than many practitioners realize.
FAQ
Why is the harmonic mean always the smallest of the three Pythagorean means?
Mathematically, the harmonic mean inverts the numbers, averages the inverses, and inverts back. Inverting large numbers makes them small and small numbers large; the averaging of inverses gives more weight to the smaller original numbers (whose inverses are larger). The re-inversion then produces a value pulled toward the smaller numbers in the dataset. This mathematical property is precisely what makes the harmonic mean appropriate for rate averaging, where slower speeds or lower ratios should receive greater weight.
Can the harmonic mean be used with negative numbers?
Generally no, or at least not without significant complications. The harmonic mean is undefined or produces counterintuitive results when values include zero or negative numbers. This limits its applicability in financial contexts where returns, growth rates, or earnings can be negative. When dealing with potentially negative ratios like P/E ratios during loss periods, alternative approaches — such as omitting negative observations or using different aggregation methods — are typically employed.
Related Terms
- Arithmetic Mean — the simple average; sum divided by count; the most familiar measure of central tendency
- Geometric Mean — the nth root of the product of n numbers; appropriate for multiplicative data like growth rates
- P/E Ratio — the price-to-earnings ratio; a valuation multiple for which the harmonic mean is the correct portfolio-level average
- Dollar-Cost Averaging — investing a fixed dollar amount at regular intervals, naturally producing a harmonic mean cost per share
- Rate — a ratio comparing two different units (e.g., miles per hour, dollars per share), typically best averaged with the harmonic mean
Related MoneyBestPal Guides
Use these related MoneyBestPal resources to compare Harmonic Mean with nearby finance, economics, investing, and business concepts.
 |
| Image: Moneybestpal.com |
In finance and economics, a harmonic mean is a statistical tool used to determine the typical rate of change over time. It is the reciprocal of a set of numbers' reciprocals' arithmetic mean.
The harmonic mean is frequently used in financial analysis to determine the average return on investment over a period of time, such as the yearly return on a mutual fund over a number of years. This is due to the harmonic mean's tendency to give greater weight to periods with lower returns, which can be advantageous in some investment types where there may be years with high returns and years with poor returns.
Think of a mutual fund, for instance, where your investment yields 5% in the first year, 10% in the second year, and 15% in the third. The harmonic mean return for this investment would be:
1 / [(1/1.05) + (1/1.10) + (1/1.15)] = 10.21%
This indicates that your average annual return would have been 10.21% if you had invested in this mutual fund for the whole three years.
The harmonic mean can be helpful in some situations, but it's crucial to remember that it's not always the best technique to gauge central tendency. It may not adequately reflect the main trend of the data and is quite sensitive to outliers. As a result, while examining data, it is crucial to take additional statistical metrics such as the arithmetic mean and median into account.
