Comments from Rupert Watson, Global Head of Economics & Dynamic Asset Allocation and the team 

Author: Rupert Watson, Global Head of Economics & DAA

Economic growth matters, because the more of it we have, the higher our incomes will be and the easier it will be for companies to generate greater profits. But, predicting economic growth is difficult. 

Over the medium to long term, increases in the number of people increase overall gross domestic product (GDP) but not necessarily GDP per capita. What really matters is productivity growth: the ability to produce more with the same or the same with less. As Nobel Laureate Paul Krugman once said, “Productivity isn’t everything, but in the long run it’s almost everything.”

So how do you increase productivity? Generally speaking, increases in education standards (skills) and greater use of effective technology should help a company or country get more out of less. Such increases tend to happen only slowly, and productivity growth tends to be fairly stable.

Some economists, however, have recently argued that AI will substantially increase productivity in the US. McKinsey recently wrote that “generative AI is poised to unleash the next wave of productivity”.  Economists estimate that advanced analytics, traditional machine learning, deep learning and AI could boost the global economy by US$17.1 trillion–US$25.6 trillion over the next couple of decades. Current global GDP is just over US$100 trillion, implying a boost of around 20%. To put it mildly, such a boost would be huge. Others argue the impact will be much more modest and largely lost in the normal ebb and flow of the global economy. So who’s right?

The truth is, I’ve no idea, although I would place myself on the optimistic side. AI is already having a significant impact on economic activity, albeit only in a small part of the economy. For example, in the US, we’ve seen a significant reduction in the number of people working in call centres, implying strong productivity growth. Elsewhere, companies are reportedly spending a lot of money trying to develop AI tools for themselves.

It’s still early days, but there are some developments/discoveries that hint that the prize could be very large indeed. Klick Labs recently reported that their AI tool and 10 seconds of voice could detect Type 2 diabetes with 89% accuracy for women and 86% accuracy for men. 

What might be possible in five years? Or 10? Or more? I don’t think anyone really knows, and it’s likely that many of the tools and discoveries that will change the world in 10 to 20 years haven’t even been thought about yet and will rely on technologies not yet invented. However, I lean toward the optimistic side, and I think AI could have a huge impact on productivity growth and, thus, economic growth, wage growth, corporate profit growth and welfare more generally.

Footnotes

1. McKinsey. Economic Potential of Generative AI.

2. Klick Health. “AI and 10 Seconds of Voice Can Screen for Diabetes, New Study in Mayo Clinic Journal Reveals,” October 18, 2023.

Author: Rupert Watson, Global Head of Economics & DAA

In terms of the size of the losses, it is too early to tell at this point, as the bonds haven’t yet been sold or allowed to mature. However, a guess/forecast can be made based on where yields are today. From quantitative easing’s (QE’s) inception up until March 2021, the BoE made a cumulative profit of £124 billion, according to data from the Office for Budget Responsibility.1 Since then, yields have risen sharply, especially at the short end.

• Since March 2010, 10-year yields are up about 3%, or down approximately 25% in price. With QE peaking at £895 billion, the losses since then might be approximately £250 billion. 

• UK GDP was approximately £2.2 trillion in 2022, so losses since March 2022 might be a bit over 10% of GDP. After taking account the gains from QEs inception till March 2022, total losses might be approximately £125 billion or 5%. 

In terms of selling bonds, there are two options that a central bank has to get its balance sheets back to normal. The first option is the slower of the two and involves not reinvesting some/all maturing bonds. The second option is quicker and involves doing option one, but also actively selling bonds. 

Whether option one or two turns out to the best from a financial perspective depends on whether the yield the Banks are selling at now is higher or lower than yields in the future. This, you can imagine, is difficult to forecast. Ex ante there is no right or wrong way of doing it. Although, there is a good argument that a central bank shouldn’t be taking a ‘punt’ on the markets and should thus exit as soon as possible. 

As of writing, the Federal Reserve is not actively selling any of the Treasuries it has bought and its balance sheet should be back to a normal size quite soon. However, the duration/maturity of the BoE’s QE gilt portfolio is a lot longer than the Fed’s. This means that without any active sells, it will remain larger than normal for many years.

The BoE is wholly owned by the government (i.e., the country). This means that any profits go to the Treasury, while the Treasury must meet any losses: the gains/losses are crystalised when the bonds mature and are not based on market prices. From QEs inception until 2022, the government was getting a nice payment each year from the BoE. However, for the next five to ten years the payments will go in the other direction, worsening the government’s annual fiscal numbers. These payments will undoubtedly make things more difficult for the government of the day, especially as budgets are likely to remain strained for several years to come. However, at a high level, is a 5% GDP loss significant? No, it is small in the broader scheme of things. 

It must also be remembered that the purpose of QE was not to make a profit. It was to boost the economy in an environment in which interest rates couldn’t be cut any further. Had the BoE not done QE then the economy would have fared much worse (say many, but not all, commentators). As a result, the loss to government caused by a reduction in tax receipts and various other payments would have been much greater than 5% of GDP. There would also have been significant societal damage caused by high unemployment and other by-products of very weak economic growth.

^Footnotes

^{1 The fiscal impact of quantitative easing and quantitative tightening}