‘It was the best of times, it was the worst of times.’ So goes the opening line of Charles Dickens’ 1859 novel, A Tale of Two Cities.

The last month has been something close to the financial markets’ equivalent. Of course, equity markets have had the best of it. ‘Plus ça change, plus c’est la même chose’, as the French would say. Nobody considers, let alone prices, the balance of risk anymore. This has been, and continues to be, a market driven by extrapolation and momentum. In this latest episode of acute equity investor hysteria, the specific beneficiaries have altered somewhat but the story remains pretty much the same. Artificial intelligence (AI) is an unassailable and all-powerful force so revolutionary and unequivocally successful and beneficial that we must all simply bow down, accept our fate and get on board the train. Apologies for the thinly veiled cynicism but for those of us who have been around the block a few times it all sounds rather familiar.

There is no arguing with the here and now. A very small number of mostly gigantic corporations are building out the capacity to operate and commercialise large language models (LLMs) and the, as-yet fairly thin, application layers which sit atop them. Hundreds of billions of pounds, dollars and yuan are being poured into data centre construction programmes, as the world’s most powerful entities vie to be the emergent winner-takes-all. That means demand for the various inputs into these energy-hungry behemoths has gone parabolic. Questions about the sustainability and ultimate profitability of these actions are for another day (or simply for the heretics). Questions about the end state of the industry likewise. Personally, I still have questions about the macroeconomic and microeconomic impact of the AI revolution. And I certainly see little evidence that the trajectory of LLMs is likely to be linear with respect to inputs. Without going deep into the weeds, these models are not clever, they are statistical. They play the odds. And while that can lead to some very impressive output, it also leads inevitably to error (hallucinations in the AI parlance). Putting more or faster processors into use does not deal with those challenges. They are fundamental. It is notable how many of the leading AI researchers are already researching different types of AI models as they admit and acknowledge the failings and limitations of LLMs. This is not me decrying LLMs as useless, rather it is assessing that their use is somewhat limited and, in my humble opinion, they are not the route to the holy grail of artificial general intelligence.

The financial aspect also raises questions for serious investors. Are we to presume that the long-term steady state for AI requires this level of demand for data centre construction? Will the companies selling the semiconductors used to train and run the models see these levels of demand long into the future? Or is it, in fact, more likely that the “buildout phase” is a demand shock and that in the end the steady state will require significantly fewer new chips each year?

What about the cost of depreciation? Buying chips for investment purposes does not hit the profit and loss statement immediately. Investing isn’t a cost – depreciation is. So, what happens when the depreciation effects of this massive build out start to hit corporate results?

Other questions abound – what does the industry for AI provision look like in the future? Does one dominant provider emerge victorious (in which case the competitors left in their wake will have presumably lost a lot of money)? If many successful providers emerge then how will they protect margins? Will that not begin to look like a commodity industry? And then there is the impact and source of the capital itself. Many of these massive hyperscalers have been hyper profitable and thus their ability to spend lavishly was huge, but most are now taking on significant debt to continue their push for global AI domination. This further raises the stakes for companies and investors alike. An even less talked about dynamic is the role of share buybacks. By any measure share buybacks have been a significant aspect of the long US equity bull market. Buying back shares, as opposed to paying dividends, has one major benefit for equity investors. By reducing the number of shares in issue, you automatically improve metrics like earnings per share (EPS), enabling higher market capitalisations as traditional valuation metrics, such as the price earnings multiple, look less stretched. Buybacks also have the rather convenient benefit of offsetting what would otherwise be significant dilution for existing investors due to lavish stock-based compensation for employees. Without buybacks, those massive stock ‘comp’ packages are going to act like equity issuance. Buyback expectations are already in decline.

It is said that equity markets climb a wall of worry (meaning that they feed on the scepticism of bearish investors who, presumably, have to close short positions or are forced to buy into the rising market in order to keep their jobs). In the modern market, you do not even really need bears to feed the market – systematic investment styles which simply buy rising prices and feed off momentum are ubiquitous.

All of which leads less-starry-eyed investors, such as me, in danger of looking the fool as we ask valid economic and investment questions while many others simply buy the vibes and look clever while doing so. Everyone is a genius in a bull market, as they say.

Starry-eyed investing and bond markets however are incongruous. Bonds have an asymmetrically negative return profile. If you hold a bond to maturity your return is near-certain, unless the issuer is unable to pay. If that happens then you lose some, if not all, of your investment. This does not lend itself to a rose-tinted view of the world. Thus, we find ourselves in a situation where bond investors have got the heebie-jeebies about inflation, stagflation, fiscal incontinence, rate hikes and political volatility – all of which are, to some degree, valid and rational concerns.

But here’s the rub. When bond yields go up, they put correcting and disciplining pressure on all of the above forces. As US political veteran James Carville so famously said back in the early 1990s, the bond market can intimidate everyone. Unless central banks interfere, of course. Which they love to do, but doing so now would risk outright rebellion. That is because it would be such an overt breach of their mandates and their fundamental raison d’être. That, in my opinion should NOT include bailing out over-levered markets every time they have a wobble. The leverage is, of course, an obvious outcome of those same central banks’ easy money policies of the past. This has led me to repeatedly describe the Federal Reserve as both arsonist and fireman.

Higher bond yields should also put downward pressure on stock valuations. Theoretically a share’s price today should equal the discounted present value of its future earnings. If bond yields rise that future stream of earnings will be worth less today because of the higher discount factor, which takes into account the yields available from bonds.

I know it is fashionable to simply react to prices. Buy the stuff that is going up and sell the stuff that is going down. That sort of thing used to be the kind of behaviour that made retail investors the “dumb money”. Now it is not only ubiquitous but celebrated by institutional investors who ought to know better. But what about taking a more medium-term view and taking into consideration not just what we know to be true to today, but what might be true tomorrow? When you do this and consider how the market is compensating you for those views, I think the opportunity set looks very different. History says that if you buy the US equity market at prevailing valuations then your 10-year return will be somewhere in the vicinity of 0% and the negativity will be relatively front loaded. Meanwhile, depending on which bond you buy, you can potentially lock in a 4-6% risk-free annual return over the next 10 years. I know there is a lot of bond-unfriendly stuff going on, but that is why yields are as high as they are. The question is not whether there are risks. There are always risks – whether you choose to pay attention to them or not. The question for investors is whether you are being compensated for those risks, and through that lens the bond market is looking interesting indeed.

‍

James Athey is Co-Manager of Marlborough's Global Bond & Global Corporate Bond funds.

‍

‍

This article is provided for general information purposes only and should not be construed as personal financial advice to invest in any fund or product. These are the investment manager’s views at the time of writing and should not be construed as investment advice. The opinions expressed are correct at time of writing and may be subject to change. Capital is at risk. The value and income from investments can go down as well as up and are not guaranteed. An investor may get back significantly less than they invest. Past performance is not a reliable indicator of current or future performance and should not be the sole factor considered when selecting funds.