The forces shaping markets, and the ones that only look like forces.

Most market commentary is about the last two weeks. The forces that actually decide what a portfolio is worth in ten years move on slower clocks: the cost of money, how much governments borrow, the price of power, who is available to work, and how concentrated the index has become. Five of those are worth understanding. The rest is weather.

Start with the cost of money. For the fifteen years after 2009, the return on safe cash was zero, and an entire generation of portfolios was built on the assumption that it would stay there. It has not. Policy rates across the rich world have settled somewhere between three and five per cent, and long bond yields have stopped falling. That one change reprices everything, because the same rate is the discount on every future cash flow, the cost of the loan against a building, and the return you earn for doing nothing. Cash is now a real alternative rather than a place to park money while you decide.

Bonds have become useful again, but not automatically. A bond only hedges an equity portfolio if it rises when equities fall. In 2022 the two fell together, the worst year for that relationship in half a century, because inflation was the problem and bonds were the thing being inflated away. Duration is a bet on rates settling lower. At current yields, we take some of that bet, in the short and middle of the curve, and we do not take it at thirty years.

Then the borrowing. American deficits are running near six to seven per cent of GDP with unemployment low, which is not a normal setting. Debt held by the public is around one hundred per cent of GDP, and interest costs have climbed past defence spending. Governments are issuing a great deal of paper, and that supply competes with everything else for capital. Japan, Britain and much of Europe are in a milder version of the same position. This does not produce a crisis on a schedule. It produces a persistent tug upward on yields, and more volatility on the weeks when an auction goes badly.

Inflation has changed shape rather than disappeared. Goods prices, which drove the surge in 2021 and 2022, have largely normalised. Services and wages have not, because labour is scarce across the rich world. Tariffs add a fixed cost to a wide range of imports, and businesses pass those on over quarters rather than days. The result is not a return to nine per cent. It is inflation that sits above target for longer than central banks would prefer, which keeps rates from falling as quickly as markets keep hoping.

The fourth force is capital spending. The largest technology companies are now spending several hundred billion dollars a year on data centres, chips and the buildings around them, a share of total business investment not seen since the railway and telecom booms. If that spending continues, the returns show up first in power generation, grid equipment, construction and cooling, long before they show up in the software it is built for. If it slows, the index feels it immediately, because the companies doing the spending are the index.

Power is the constraint most people still underestimate. Electricity demand in America fell for two decades and is now rising, while new generation takes years to permit and connect. Interconnection queues in several markets run five years or longer. Cheap, reliable electricity is becoming a competitive advantage again, a phrase nobody in this industry used for thirty years, and the companies that own generation or the equipment to move it are being repriced accordingly.

Trade has stopped growing faster than the world economy, which reverses a forty-year trend. Tariffs, export controls and the deliberate duplication of supply chains for semiconductors, batteries and pharmaceuticals all point the same way. Duplication is expensive: two factories making the same thing cost more than one, and someone pays for the spare capacity. That someone is the consumer, through higher prices, and the shareholder, through lower margins at the companies doing the duplicating. It is also why parts of heavy industry are worth more than their recent earnings suggest.

Demographics sit underneath all of it. The largest cohorts in the rich world are leaving the workforce, and immigration policy has become restrictive in several of the countries that most needed it. Labour scarcity is no longer cyclical. A tight labour market combined with tariffs is an inflationary combination, and it limits how far central banks can cut even when growth disappoints.

The structure of the market itself has changed. The number of listed companies in America has roughly halved since the mid-nineties while the economy has doubled, because companies stay private longer and arrive on the exchange older and larger. Public indices therefore miss a slice of the growth, and investors have pushed into private equity and private credit to compensate, often without pricing the illiquidity honestly. Meanwhile passive funds hold more American equity than active ones, so every new dollar buys more of whatever is already largest. Ten companies are now close to forty per cent of the S&P 500. A portfolio described as diversified is, in the way that matters, a bet on those ten.

Outside the market, central banks have been buying gold at the fastest rate in decades and shifting reserves away from a single currency. That is not a prediction of collapse. It is a slow change in what the world treats as a safe asset, and for a family holding money in three currencies it means the assumption that the dollar asset is always the safest one deserves a re-examination every few years rather than a lifetime.

Now the part that is noise. A single earnings miss, an election, one sentence from a central banker, a strike, a resignation: these produce movement and almost no lasting change in value. The forces above move over years. A portfolio rebuilt around each quarter's headlines will be rebuilt six times a decade, will pay the tax and the spread each time, and will underperform the one built around the slower forces. Attention is the scarcest input in this business, and most of it is spent on the wrong clock.

So what do we actually do about it. We hold three to five years of spending in Treasury bills and short bonds, so that no market decides when a client sells. We keep bond duration short and take credit risk deliberately, in modest size, rather than reaching for yield at thirty years. We own some of the picks and shovels of the capital spending cycle, in power, grid and industrial equipment, and we care about the price we pay for them. We hold more of the world outside the American index than we did five years ago, which is uncomfortable and correct. And we spend effort on the two things a client can control, cost and tax, because the outcome of those is not a forecast.

What we do not know is worth stating. We do not know when the spending cycle turns, whether inflation settles at two and a half or at three, or what a fractured trading system does to corporate margins over ten years. Anyone who tells you otherwise is guessing with your money. The plan has to survive being wrong about any single one of these, which is why it is built out of several, and why the position sizes are set so that no single error is fatal.

The conclusion is not dramatic. The portfolio that works for the next decade looks different from the one that worked for the last fifteen: more cash, shorter bonds, more of the world beyond one index, and fewer holdings that cannot be sold on a Tuesday. None of that requires a view on next quarter, and all of it requires patience, which remains the only edge that does not get arbitraged away.