A heartbeat
that trends up

A share is a legal slice of ownership in a company. Not a bet on a ticker, not a lottery number, not a number that goes up if enough people believe in it. If a company has ten million shares and you hold one hundred, you own a hundred ten-millionths of everything it owns and everything it earns.

That ownership carries real rights. You are entitled to a proportional claim on profits distributed as dividends, a vote at the annual meeting, and whatever is left over if the company is wound up, which in practice is usually nothing, because shareholders are last in the queue behind lenders, suppliers and staff.

This matters more than it sounds. If you hold a share, you are not hoping someone pays more for it later. You are entitled to your slice of the profits that company earns for as long as you hold it. The price is what someone will pay you today for that future stream of earnings. Everything else in investing is an argument about what that stream is worth.

You own a fraction of a business, not a number on a screen.

Compounding is what happens when returns start earning returns. It is the single most powerful thing available to a private investor and the one that feels the most like nothing is happening for the longest time.

A thousand pounds a month at eight percent for ten years gives you about £180,000, of which roughly £60,000 is growth. Run the same thing for thirty years and you get around £1.41 million, of which £1.05 million is growth. You tripled the time and increased the growth roughly seventeen-fold. The extra came almost entirely from the last decade, when the returns on the pile were larger than anything you could have added yourself.

The uncomfortable implication is that compounding is mostly a test of whether you can leave something alone. The curve only bends if the money stays in through the years where it appears to be doing nothing, and through the crashes where it appears to be doing harm.

Time does the heavy lifting. Your job is mostly to not interrupt it.

A company can create new shares out of thin air and sell them for cash. When it does, the business gets bigger but your slice of it gets smaller. That is dilution.

Say a company has one hundred million shares and earns one hundred million pounds, so a pound of earnings per share. It issues twenty million new shares to fund an expansion. Even if profits are unchanged, earnings per share drops to about 83p. You own the same number of shares and a smaller claim on the same profits.

This is not automatically bad. If the cash raised earns a better return than the business already makes, every shareholder ends up better off despite the smaller slice. What should make you uneasy is a company that issues shares repeatedly to fund ordinary operations, because that is not investment, it is a business quietly asking its owners to top up the tank.

Watch the share count over several years, not just the earnings. A company growing profits ten percent a year while growing its share count twelve percent a year is making its owners poorer while its headlines get better.

Check whether the pie grew faster than the number of slices.

Capital expenditure, or capex, is money a company spends on things it will use for years. Factories, data centres, machinery, fleet. It is different from operating expenses, which are the costs of this year's trading, because capex buys an asset that keeps producing long after the cheque clears.

The accounting treatment is where it gets interesting. Operating costs come straight off this year's profits. Capex does not: the asset goes on the balance sheet and is charged against profits gradually as depreciation over its useful life. So a company spending enormously on new capacity can look more profitable in the short term than its cash position deserves.

This is why free cash flow matters. Profit can look healthy while the company is haemorrhaging cash into construction. When you see capital expenditure running well ahead of operating cash flow, the company is funding its future from its balance sheet or from lenders, and it had better be right about the demand it is building for.

Profit is an opinion. Cash going out of the door is a fact.

Price to earnings is the share price divided by earnings per share. If a share costs £150 and the company earns £5 a share, the P/E is thirty. The plain English reading: you are paying thirty years of current earnings for one share of that business.

A high P/E is not the same as expensive. It means the market expects earnings to grow, and is willing to pay in advance for that growth. A low P/E is not the same as cheap. It often means the market expects earnings to fall, and it is frequently right.

The ratio is only useful in comparison, and only between similar things. A supermarket on a P/E of thirty and a software company on a P/E of thirty are not equally valued, because one of them can grow revenue without building more shops. Compare a company to its own history and to its direct competitors, never to the market average.

Two traps. Trailing P/E uses last year's earnings and is backward-looking. Forward P/E uses a forecast, which means it is only as good as the forecast.

A P/E is a question about expectations, not an answer about value.

PEG takes the obvious weakness of the P/E and patches it. Divide the P/E by the expected annual earnings growth rate and you get a number that tells you what you are paying per unit of growth.

A company on a P/E of thirty growing earnings at thirty percent a year has a PEG of 1.0. A company on a P/E of fifteen growing at five percent has a PEG of 3.0. On the raw P/E the second looks half the price. On growth-adjusted terms it is three times as expensive.

The rough convention, which comes from Peter Lynch, is that a PEG below one is interesting and a PEG well above two is demanding. Treat those as rules of thumb rather than thresholds, because the entire calculation rests on a growth forecast, and growth forecasts are the most optimistic numbers in finance.

PEG also breaks completely on companies with negligible or negative earnings growth, which is most of the cyclical market at the bottom of a cycle. When the denominator approaches zero, the ratio stops meaning anything.

Cheap and slow can cost more than dear and fast.

The single most common mistake a new investor makes is assuming good news moves a share up. It does not. What moves a share is the gap between what happens and what was already expected.

A company can report record revenue, its best quarter in history, and fall eight percent on the day. That is not irrational. The price already contained an expectation of a record quarter, and the record delivered was slightly smaller than the one anticipated. The news was good. The surprise was bad.

This is what people mean by priced in. At any moment the price reflects the market's collective forecast of the future. Buying a company because it has a great product or a strong brand only works if the market has not already noticed, and with large companies it always has.

The practical consequence is that your edge cannot be knowing a company is good. Everyone knows. Your edge, if you have one, is either a different view of how good it will be, or a time horizon long enough that the next quarter stops mattering.

You are not betting on the news. You are betting against the forecast.

A pound today is worth more than a pound next year, because today's pound can be invested and next year's cannot. The rate at which you shrink future money to compare it with money now is the discount rate.

This is the machinery underneath every valuation. A company is theoretically worth all the cash it will ever produce, with each future year's cash discounted back to today. Cash arriving in twenty years is worth a fraction of the same sum arriving tomorrow.

The discount rate is anchored to interest rates, which is why rates moving matters so much and so violently. When the risk-free rate rises from two percent to five percent, every future pound in every valuation model is worth less. Companies whose profits sit mostly far in the future, which is to say growth companies, take the heaviest hit, because they have the most distant cash to discount.

That is the real reason a hawkish Fed can knock ten percent off a technology index in a week without a single company's actual business changing. The business did not change. The value of its future changed.

Interest rates are gravity. Everything priced on the future feels them.

Borrow where money is cheap, invest where it pays more, pocket the difference. That is the carry trade, and enormous sums of global capital are doing it at any moment.

The classic version borrows in Japanese yen at near zero and buys assets denominated in a higher yielding currency. As long as the interest differential holds and exchange rates stay put, it is close to free money, which is exactly why so much leverage piles into it.

The danger is the unwind. The trade depends on the funding currency staying weak. If the yen suddenly strengthens, the borrowed amount grows in real terms, positions are forced shut, assets are sold to repay loans, and the selling pushes the funding currency higher still. It feeds on itself.

You will probably never place a carry trade. You should still know it exists, because when one unwinds you will see assets with no obvious connection to Japan fall hard for no company-specific reason at all. That is not your thesis breaking. That is somebody else's margin call.

Sometimes the selling has nothing to do with the thing being sold.

The Phillips Curve is the proposition that unemployment and inflation trade off against each other. Low unemployment means workers have bargaining power, wages rise, businesses raise prices to cover them, inflation follows. High unemployment does the reverse.

This is the intellectual foundation for how central banks behave. When a central bank raises rates to cool inflation, it is deliberately trying to slow the economy enough to loosen the labour market. That is the polite way of saying it is trying to create some unemployment.

It is also, at this point, genuinely contested. The 1970s produced stagflation, high unemployment alongside high inflation, which the simple curve says cannot happen. The 2010s produced very low unemployment with stubbornly absent inflation. Economists have patched it with expectations and a long-run vertical version, but the honest position is that the relationship is unstable.

Know it anyway. Central bankers still reason with it, and central bankers set the discount rate that prices every asset you own.

A flawed model still moves markets if the people with the rates believe it.

Classical theory assumes investors are rational. Forty years of evidence says they are reliably, predictably irrational in ways that can be catalogued.

Loss aversion is the big one. Losing a hundred pounds hurts roughly twice as much as gaining a hundred feels good. This is why people hold losers far too long, refusing to make a paper loss real, while selling winners early to lock in the pleasant feeling.

Anchoring fixes you to an irrelevant number, usually the price you paid. The market has no knowledge of your entry price and no interest in returning to it. Confirmation bias sends you looking for reasons you were right. Recency bias makes the last six months feel like a permanent state of the world. Overconfidence convinces you that your gains came from skill and your losses from bad luck.

The defence is not becoming rational, because you will not. It is building rules in advance, when nothing is happening, and then following them in the moment when everything is. That is the entire reason I write my reasoning down publicly before I act on it.

You cannot remove the bias. You can make a decision before it shows up.

If you sell an investment outside an ISA or pension for more than you paid, the profit is a capital gain and it is taxable. For the 2026/27 tax year the annual exempt amount is £3,000. Gains beyond that are charged at eighteen percent within your basic rate band and twenty-four percent above it.

Three things catch people out. First, the allowance cannot be carried forward. Unused on 5 April, it is gone. Second, losses in the same tax year are set against gains in that year before the allowance is applied, so realising a loss in a year with small gains can waste the allowance entirely. Losses brought forward from earlier years behave differently and only reduce gains down to the allowance. Third, losses must be formally claimed, generally within four years, or you cannot use them at all.

Then there is the thirty day rule. Sell a holding and buy it back within thirty days and the disposal is matched against the repurchase, so the gain or loss is not crystallised as you intended. It exists specifically to stop people resetting their base cost for free.

The practical upshot for a long-term investor is that selling is expensive. A rebalance that looks tidy on a chart can cost you a quarter of the gain.

The allowance expires every April. The tax on selling never does.

HMRC does not treat cryptoassets as currency. For almost every private individual they are property, and the same capital gains regime applies as for shares, with the same £3,000 annual exempt amount and the same eighteen and twenty-four percent rates.

The part people miss is what counts as a disposal. Selling for pounds is obvious. Swapping one token for another is also a disposal, and you owe tax on the gain in the first token measured in sterling at the moment of the swap, even though no pound ever touched your bank account. Spending crypto on goods is a disposal. Gifting it to anyone other than your spouse is a disposal.

Tokens of the same type are pooled under section 104 rules, with the same thirty day matching rule as shares. Mining, staking and airdrop rewards are usually income at the point of receipt and then have a separate capital gain when eventually sold.

The record keeping is genuinely onerous, and the obligation sits with you rather than the exchange. If you have traded actively across several tokens, this is an area where paying an accountant is cheaper than the alternative.

Swapping one token for another is a taxable event, even with no cash involved.

An initial public offering is a private company selling shares to the public for the first time. It hires investment banks as underwriters, who value the business, gauge institutional demand through a bookbuilding process, and set a price.

It is worth being clear-eyed about who the process serves. The underwriters' clients are the company selling and the institutions buying at the offer price. Retail investors generally cannot buy at the offer price at all; they buy on the first day of trading, at whatever the market opens at, which is frequently well above what institutions paid that morning.

There is also a lock-up period, typically ninety to one hundred and eighty days, during which insiders and early backers cannot sell. When it expires, a large volume of shares can become available at once, and the share price often takes that badly.

None of which says avoid IPOs. It says that the first day is the single worst-informed moment to form a view on a company that has no public trading history, no track record of guidance, and a set of sellers who know considerably more than you do.

Being early to a listing is not the same as being early to the company.