You will be able to judge how well management has used the cash the business produced.
At Larkspur's last annual meeting, a shareholder asked why the company was sitting on S$60 million of cash. The chief executive said the board was "keeping its options open for growth". Marcus had heard the same words from other boards before acquisitions that went badly. Over a decade, what management does with the cash a business produces can matter as much as the business itself, and the record is in the cash flow statements for anyone who adds it up.
Figures are made up and in S$ millions.
Every dollar of cash a company produces ends up in one of five places: reinvestment in the existing business through machines, working capital or research; the purchase of another business; repaying debt; dividends; or buying back the company's own shares. Keeping the cash in the bank is a sixth, temporary choice that sooner or later turns into one of the five.
None of these is good or bad in itself, and each is judged the same way: did it earn more than the cost of capital, about 9% for Larkspur from lesson 8.2, Cost of capital: equity, debt and a WACC you can defend? Reinvestment and acquisitions earn whatever the assets they buy go on to earn. Debt repayment earns the interest saved, and lowers risk. Dividends hand the choice to shareholders. Buybacks earn the gap between the price paid and the value of the shares.
The test is a simple table. Take ten years of cash flow statements, add up where the cash came from, then add up where it went, by category.
When Marcus did this for Larkspur, operating cash flow over ten years totalled 520 and the property sale brought in another 10, so 530 came in. It went out as 290 of capital spending, 55 for the coatings business, 140 of dividends and 20 of net debt repayment, with the remaining 25 added to cash. Both columns total 530, the same discipline as the cash check in lesson 7.7, Enter five years of history into your model.
Now judge each line against the returns that followed.
Reinvestment, at 290, was by far the largest use. Almost all of it went into the parts business, which lesson 9.4, Competitive position: Five Forces and moats, tested against the numbers, found earning about 19% on its capital. That's well above 9%, so the bulk of Larkspur's spending created value.
The acquisition cost 55, of which 30 remains on the balance sheet as goodwill. In its best year since the deal, coatings earned about 6% on that price, or 3.3 a year after tax. A stream of 3.3 a year for ever, valued at 9%, is worth about 36.7, so measured on its best year the deal still lost about 18 of the 55 paid. With coatings near break-even in FY5, the real loss is larger. Lesson 6.5, Management incentives, ownership and related-party deals, found the chief executive's bonus tied to revenue and EBITDA, both of which the deal raised on day one.
Dividends of 140 went to shareholders, who could reinvest elsewhere. Larkspur paid 18 in FY5, about 43% of underlying profit of 42. A company with no investments that clear its cost of capital should pay out more. Larkspur did have such investments, in its parts business, and still paid a steady dividend, so this looks like a balanced split.
Debt repayment of 20 lowered risk and saved interest. With net debt now about a quarter of a year's EBITDA, more repayment would add little.
Larkspur made one acquisition. Some companies make one every year, and their record needs a sharper test, because a single good deal can hide several poor ones.
Track goodwill and intangible assets against operating profit over five or ten years. Take a made-up serial acquirer whose goodwill rose from 100 to 400 over five years while operating profit went from 50 to 70. It has paid at least 300 above the book value of what it bought, and its profit has risen by 20. Even if every extra dollar of profit came from the acquisitions, that's under 7% a year before tax on the premium alone, below most companies' cost of capital. Goodwill climbing far faster than profit is the signature of a company buying revenue rather than returns.
Look out for one more sign: impairments that arrive a few years after each deal. They're the company admitting, late, that it paid too much.
Lesson 5.5, Corporate actions: rights issues, placements, buybacks and splits, showed that a buyback moves value between the holders who sell and the holders who stay. Now Marcus has a value to test it against.
His model from lesson 7.8, Build the forecast and balance the model, has Larkspur's cash rising to about 136 by FY10, so the board's "options" question will only get louder. Suppose it bought back 30 million shares, a tenth of the total, at today's price of S$1.60. It would spend 48. On Marcus's DCF, the equity is worth about 538.6, so the remaining holders would own about 490.6 across 270 million shares: about S$1.82 each, a modest gain on the S$1.80 they were worth before. If the shares had run up to S$2.10 first and the company bought the same number, it would spend 63, leaving about S$1.76 a share. The same buyback, at a price above value, makes the holders who stay poorer.
So when a company announces a buyback, ask what a share is worth and what the company is paying for it, two questions the announcement never answers and your valuation does.
Over ten years a good allocator shows most cash going into investments that earned well above the cost of capital, acquisitions that kept earning after the first year, a dividend set at a level the business can sustain through a downturn, and buybacks concentrated in years when the shares were cheap. Larkspur's record was mixed: good reinvestment, one value-destroying acquisition, sensible dividends, and a growing cash pile with no stated plan.
Gather the cash flow statements for your own company for as many of the last ten years as you can find, ready to total each use of cash.
Total your company's uses of cash over the last ten years by category and judge which ones earned their cost of capital.
Junxiong-WFG Organisation is an authorised representative of AIA Financial Advisers Private Limited (Reg. No. 201715016G).