Realised Yield Approach in Estimating Cost of Equity

When a company wants to know what return its shareholders expect, it has two broad choices: build a model around assumptions, or look at what actually happened. The realised yield approach takes the second route. Instead of forecasting future dividends or estimating a market risk premium, it simply asks what shareholders have earned in the past and uses that number as the best available estimate of the cost of equity.

What the realised yield approach actually measures 🔗

The realised yield approach estimates the cost of equity by calculating the actual returns investors received from a stock over a series of past periods, then averaging them. This average, expressed as a percentage, becomes the proxy for Ke, the cost of equity.

The logic rests on a simple idea from finance theory: over a sufficiently long stretch of time, unexpected gains and losses should roughly cancel each other out, leaving the average realised return as a reasonable stand-in for the return investors originally required. This is essentially the same reasoning that underpins the “average realised returns” school of thought discussed in classic cost of equity research, where actual historical stock performance is treated as a proxy for the return investors expected when they bought in.

Unlike the dividend growth model, which needs an assumed future growth rate, or the Capital Asset Pricing Model, which needs a risk-free rate, a beta, and a market risk premium, the realised yield approach needs none of these forecasts. It works entirely with numbers that have already happened.

Why go backward to estimate something forward-looking 🔗

Cost of equity is, by definition, a forward-looking figure. It represents the return a company must offer to keep attracting equity investors going forward. So why calculate it using historical data?

The answer is practicality. Forecasting future dividends, growth rates, or risk premiums involves a fair amount of guesswork. Historical returns, by contrast, are observable and verifiable. If a company has traded on the stock exchange for several years and has a reasonably stable history, its past yield pattern can offer a grounded, data-backed starting point rather than a purely theoretical one. This is precisely why the approach appears in the cost of capital curriculum studied by commerce and finance students, positioned alongside the dividend growth model and CAPM as one of the standard ways to arrive at Ke.

How the calculation works, step by step 🔗

The mechanics of the realised yield approach are straightforward once broken into stages.

Step 1: Collect historical price and dividend data 🔗

The first requirement is a data set covering several years, typically four to five, containing the beginning share price, ending share price, and dividend paid for each year. The longer and more stable the trading history, the more reliable the estimate tends to be.

Step 2: Calculate the yield for each year 🔗

For every year in the data set, the yield is calculated using this formula:

Yield = (Dividend received + Ending price − Beginning price) ÷ Beginning price

This captures both sources of shareholder return: the dividend income and the capital gain or loss from the change in share price.

Step 3: Average the yields 🔗

Once each year’s yield is calculated, the final step is averaging them to arrive at a single figure representing the realised yield. This is where the approach involves a genuine choice, and it matters.

In this illustrative set, the simple average of the four yields comes to roughly 9.2 percent, and this figure would serve as the estimated cost of equity under the realised yield approach.

Arithmetic mean or geometric mean: it changes the answer 🔗

The choice between an arithmetic average and a geometric average is not a minor technical detail. It can shift the final estimate meaningfully, especially when returns are volatile.

The arithmetic mean simply adds up the yearly yields and divides by the number of years. It is easy to calculate but tends to overstate the return an investor would have actually experienced if they held the investment across the entire period, particularly when returns swing between large gains and losses.

The geometric mean, by contrast, compounds the yearly returns and extracts the equivalent constant annual rate. This is closer to what an investor’s money actually grew at over the holding period. Research on this exact trade-off notes that when stock prices show meaningful volatility, the geometric mean tends to be the better estimator of expected future returns, since it reflects compounding rather than a simple average of ups and downs.

As a rule of thumb: if the yearly yields in your data set are fairly close to each other, both methods will give similar answers. If the company’s stock has been through wide swings, such as a sharp market correction followed by a strong recovery, the geometric mean will produce a more conservative and arguably more realistic estimate of Ke.

Advantages that make this approach attractive 🔗

The realised yield approach has a few genuine strengths that explain why it remains part of standard financial management coursework.

• Grounded in real data: It does not depend on assumptions about future growth rates, risk premiums, or beta coefficients. Every input has already been observed in the market.
• Simple to compute: The formula involves basic arithmetic, making it accessible even without advanced financial modelling tools.
• Transparent and verifiable: Since the inputs are historical share prices and dividends, anyone can cross-check the calculation using publicly available data.
• Useful for established, actively traded companies: Firms with several years of consistent trading history and dividend payments are well suited to this method.

Where the approach falls short 🔗

The same feature that makes this method appealing, its reliance on the past, is also its biggest weakness. A company’s future is not obligated to resemble its history.

Regulators take this limitation seriously enough to mandate warnings about it. In the mutual fund industry, for instance, disclosures are required stating that a fund’s past performance does not necessarily predict its future results. The same caution applies when using a company’s historical stock yield to estimate its future cost of equity.

A few specific problems are worth flagging:

• Earnings and prices are rarely stable: Corporate earnings fluctuate with business cycles, competition, and policy changes, so a return pattern from five years ago may not hold going forward.
• Sensitive to the time period chosen: Picking a different starting or ending year for the data set can swing the result considerably, since a single strong or weak year can dominate a short average. Analysis of index-linked return calculations has shown how the choice of start and end dates for a compounded return figure can present a misleadingly rosy or gloomy picture depending on which window is selected.
• Not usable for new or thinly traded companies: Without a meaningful trading history, there is no realised yield data to work with in the first place.
• Ignores changing risk profiles: A company that has become more or less risky over time may not have a cost of equity that matches its old yield pattern.

How it fits alongside other cost of equity methods 🔗

Financial managers rarely rely on a single approach in isolation. The realised yield approach is typically used alongside the dividend growth model, which relies on projected future dividends and an assumed growth rate, and the Capital Asset Pricing Model, which links the cost of equity to systematic risk through beta. Where the realised yield approach shines is as a sanity check: if a CAPM-based estimate comes out far higher or lower than what the stock has actually yielded historically, that gap is worth investigating before finalising a cost of capital figure for a project or valuation.

Putting it into practice 🔗

For a student solving a numerical problem, the process is mechanical: build the yearly yield table, decide whether the question calls for an arithmetic or geometric average, and compute accordingly. For a working analyst, the harder judgment call is choosing how many years of data to use and whether the company’s recent history is even representative of its likely future. A five-year average calculated right after an unusual boom year, or right after a crash, can distort the picture unless it is read with context.

What do you think? If a company’s stock delivered a strong realised yield during a period of unusually high market growth, would you trust that number as its future cost of equity? And when a firm’s earnings have been particularly volatile, does an arithmetic average or a geometric average give management a more honest picture of what shareholders actually expect?

References

1. https://www.rba.gov.au/publications/rdp/1991/9107/cost-of-equity-theory.html
2. https://live.icai.org/bos/vcc/pdf/Ch_4_Cost_of_Capital_ICAI_for_Nov21.pdf
3. https://www.sciencedirect.com/topics/economics-econometrics-and-finance/cost-of-equity
4. https://www.investor.gov/introduction-investing/investing-basics/glossary/mutual-funds-past-performance
5. https://www.business-standard.com/amp/article/pf/a-better-tool-to-evaluate-funds-116021300843_1.html