Python for programmers, prompts for data analysts.

For the last four parts, the adjusted close has been the hero of this series. We rebuilt it from raw prices, splits and dividends. We saw that dividends made up 40% of Coca-Cola's ten-year
return. And we saw where the adjusted close misleads a backtest.

At the end of Part 38 I promised to look at the other way a company returns cash to its shareholders. A dividend puts money in your account. A buyback doesn't: the company buys its own
shares on the market, and fewer shares are left. Apple pays a small dividend, but it spends far more on buybacks.

So the question for this part: how much did Apple return through buybacks over ten years, and why does none of it show up in the adjusted close?

Want to try this yourself? The EODHD MCP Server is what pulled the data in this article — connect it to Claude and ask your own questions
(affiliate link).

New here? The Setup covers everything you need before your first prompt.

Two ways to return cash

Apple's fiscal year ends in late September, so ten fiscal years, 2017 to 2026, run from October 2016 to September 2026. The numbers come from Apple's cash flow statements.

🟧 PROMPT

Use the eodhd MCP to get Apple's cash flow statements for fiscal years 2017 to 2026. How much did Apple spend on buying back its own shares, how much on dividends, and how does that compare with its free cash flow? 

Over the ten fiscal years:

10 years in total
Share buybacks $769 billion
Dividends $145 billion
Free cash flow $895 billion

Apple spent 5.3 times as much on buybacks as on dividends. The gap was there every single year: from 2.5 times in fiscal 2017 to 6.2 times in fiscal 2024, when buybacks peaked at $94.9
billion. Together, buybacks and dividends came to 102% of free cash flow. Apple returned slightly more than the cash its business generated.

In Part 36, dividends added 47.5 percentage points to Apple's total return. Buybacks moved five times as much money, yet they appear nowhere in that calculation. Dividends were the visible
part of Apple's cash returns. Buybacks were most of it.

So where did $769 billion go? Not into anyone's account. It went into the market, to buy shares from whoever wanted to sell.

Fewer shares, bigger slices

A buyback doesn't send cash to the shareholders who keep their shares. It changes what each of their shares stands for.

🟧 PROMPT

How did the number of Apple shares outstanding change from September 2016 to September 2026? What does that mean for someone who held the same shares the whole time? 

In September 2016, Apple had 21.6 billion shares outstanding. In September 2026, it had 14.6 billion, 32% fewer. Both numbers are in today's shares, after the 4-for-1 split of 2020.

Someone who held the same shares the whole time didn't buy or sell anything, but their slice of the company grew. Each share used to stand for 1/21.6 billionth of Apple. Now it stands
for 1/14.6 billionth. That's a 47% bigger slice of the same company: of its profits, its cash and its future dividends.

Apple's shares outstanding, quarterly, split-adjusted. Snapshot as of November 18, 2026.

The line goes down almost every quarter. There's no single day when it happens, no ex-date, no jump in the price. Apple buys a little in the market week after week, and the share count
shrinks quietly.

That's the key difference from a dividend. A dividend is an event: on one day, cash leaves the company and the price drops by about the same amount. A buyback has no such day. It works
through the count, not through the price.

Profit, and profit per share

If each share stands for a bigger slice, each share should also earn more than the company's profit growth alone would suggest. That's easy to check.

🟧 PROMPT

Compare Apple's growth from fiscal 2016 to fiscal 2026 in revenue, net income, and net income per share. For the per-share number, divide net income by the average share count during each year. 

From fiscal 2016 to fiscal 2026:

Fiscal 2016 Fiscal 2026 Growth
Revenue $215.6 billion $478.8 billion +122%
Net income $45.7 billion $132.2 billion +189%
Net income per share $2.08 $8.98 +332%

The per-share figure grew much faster than the profit itself. Net income almost tripled, but the average number of shares fell from 22.0 billion to 14.7 billion. So each share's part of
the profit more than quadrupled.

The $2.08 for fiscal 2016 matches Apple's reported earnings per share of $8.31, divided by four for the 2020 split. So the simple method works.

Here's where the buybacks went. They didn't make Apple earn more. They split the same profit among fewer shares. And a share price tends to follow what each share earns over time, so most
of the buyback effect ends up in the price itself, mixed in with everything else that moves it.

Why the adjusted close can't see it

The adjusted close corrects the price for events that move it mechanically on a single day. A split doubles or quadruples the number of shares overnight, so the price falls by the same
factor. A dividend takes cash out of the company on the ex-date, so the price drops by about that amount. Both leave a mark you can point to in the raw prices, and both get undone in the
adjusted series.

🟧 PROMPT

Why does the adjusted close account for dividends but not for buybacks? And how big were Apple's buybacks compared with its dividends, relative to its market value, in fiscal 2026? 

Because a buyback doesn't push the price down on any day. When Apple buys a share at the market price, the company has less cash, but there is also one share fewer. Each remaining share
still stands for the same value as before. Nothing needs undoing, so the adjusted close leaves it alone.

The effect comes later and slowly: more profit per share, as the last section showed. That ends up in the raw price, mixed in with everything else that moves it. So the adjusted close
does include the buyback effect. It just can't show it as a separate line.

In fiscal 2026, Apple spent $86.4 billion on buybacks and $15.7 billion on dividends. At a market value of about $5.0 trillion in October 2026, that's about 1.7% of the company bought
back in a year, against a dividend yield of about 0.3%.

So "dividends added about 6% of Apple's total return", from Part 37, is true but incomplete. It counts only the cash that arrived in an account. Most of what Apple returned to its
shareholders came the other way, as a slowly growing slice of the company.

That matters when you compare companies. Dividends, like Coca-Cola's in Part 37, show up in the adjusted close as a separate effect. Buybacks, Apple's main way of returning cash, are
hidden inside the price. Two companies can return the same amount of cash, and only the one that pays it as dividends will look like it did.

A word of honesty

Not every buyback dollar shrinks the share count. Over the same ten years, Apple paid $89 billion in stock-based compensation, mostly as new shares for employees. Part of the buybacks
simply bought those shares back. The 32% drop in the share count is the net result, after both.

A buyback only helps the shareholders who stay if the shares are bought at a sensible price. Paying too much for its own shares is a way for a company to waste cash, and nothing in this
data tells us whether Apple paid too much. That question needs a valuation, not a cash flow statement.

The per-share profit here is my own simple calculation: net income divided by the average of the quarter-end share counts. Apple reports a diluted figure based on a weighted average. For
fiscal 2016 the two agree, $2.08 against Apple's $8.31 divided by four. For other years they may differ slightly.

The market value of about $5.0 trillion uses the share count at the end of September 2026 and the price from mid-October. The yields that depend on it are rounded for that reason.

Taxes treat the two kinds of return differently. A dividend is taxed when it's paid; for a Czech investor in a US stock, part of it is usually withheld before it even arrives, as Part 37
showed. A buyback creates no taxable payment for the shareholders who keep their shares; the gain is taxed only when they sell. Not investment advice.

Final Thoughts

This is the thirty-ninth article in the series Unlock Real-Time Market Intelligence with EODHD and Claude, and the sixth about the adjusted close, after Parts 33, 35, 36, 37 and 38. Those parts showed what the adjusted
close adds and where it misleads. This one showed what it can't separate at all.

Here's what came out: over ten years, Apple returned $769 billion through buybacks and $145 billion through dividends. The dividends appear in the adjusted close as a clear effect. The
buybacks don't, even though they were five times larger. They worked through the share count, which fell by 32%, and through profit per share, which grew 332% while profit itself grew 189%.

So a chart of total return tells you how much a shareholder earned. It doesn't tell you how the company paid for it. For that, you need the cash flow statement and the share count, and
both are one prompt away.

Next time I want to put the three ways of returning cash side by side for a handful of large companies: dividends, buybacks and paying down debt. Together, Meb Faber called them
shareholder yield. Some companies pay almost everything as dividends, some almost everything as buybacks, and a price chart can't tell them apart.

If this made you curious, the MCP Server is free to try — The Setup walks you through it.

← Part 38: Prices Nobody Saw: The Lookahead Hidden in Adjusted Close