I’ve been reading Martin Wolf’s The Crisis of Democratic Capitalism this week in order to better understand the economic transition which is currently underway, beyond the enticing yet slightly vacuous claim that we are leaving neoliberalism and entering something worse. The factor I’m particularly interested in understanding, given its rhetorical and economic significance to the uptake of AI technologies, is productivity. In the book Wolf points out how stagnant productivity creates a ‘zero-sum economy’:
In a country with fast increases in productivity everybody will get better off, unless inequality rises very quickly. But in A country with stagnant productivity, such as Italy over the last two decades or the UK over the last one and a half decades, the standard of living can rise for some only if the standard of living for others falls. This then becomes a zero sum economy. If A wins, B through C must lose.
This is compounded if I understand correctly by the politics of inflation i.e. the struggle over who bears the cost of the decreasing value of money. As ChatGPT helpfully summarised it for me with regards to these distributional struggles: “For instance, inflation erodes the purchasing power of fixed-income individuals, such as retirees living on fixed pensions. On the other hand, individuals with significant assets or investments may benefit from inflation as the value of their assets increases”. But it was also preceded by a declining share of productivity gains going to labour, though the trend in the UK is less sharp than than in the US:
The report “Have productivity and pay decoupled in the UK?” published by the London School of Economics’ Programme on Innovation and Diffusion (POID) shows that between 1981 and 2019, prior to the Covid-19 hit, productivity rose by 87 per cent but median employee wages only rose by 62 per cent: a 25 percentage point “overall decoupling” between productivity growth and median wage growth.
https://www.lse.ac.uk/News/Latest-news-from-LSE/2021/k-November-21/Wages-of-typical-UK-employee-have-become-decoupled-from-productivity
With the routine caveat this is me thinking-out-loud while learning this subject matter, my understanding is therefore we have three interconnected processes:
- Declining share of productivity gains going to labour (i.e. class politics of neoliberalism of Harvey’s sense)
- Declining productivity itself generating intra-class conflict over resources (Wolf suggests pre-crisis productivity growth as partly a facsimile generated by financialisation, though it has been stagnant across all high income countries)
- The return of inflation adding in a new layer of distributional struggles
Under these conditions it is no longer possible to mobilise cheapness as a strategy, in the sense described by Patel and Moore: “Cheap things are thus not really things at all—but rather strategies adopted by capitalism to survive and manage crises, gambits made to appear as real and independent entities by the original sin of cheap nature”. The legitimating mechanisms which papered over divisions cease to be sustainable and the underlying antagonisms now exist on the surface, hence the Conservative’s desperation to entrench ‘small boats’ as a doorstep issue and their likely far right-drift after they lose the next election. Not only are people getting poorer in real terms at an astonishing speed, transforming the parameters of electoral politics in the process (e.g. Liz Truss’s government, the logical expression of Tufton Street, spending billions on an energy price subsidy) but there’s a conflict between people and between people & organisations over how the costs of this are distributed.
This is the context in which generative AI, which for all the self-serving claims made about it at this crucial point in a hype cycle can nonetheless do some remarkable things, is hitting the economy. I’m increasingly convinced this is the most salient factor in what comes next, even though it tends to be mystified by the existential risk discourse emanating from the grand old men of AI which is eating up public attention. If organisations feel under intensifying economic pressure, with cheap energy and cheap capital (seemingly) becoming a thing of the past as we enter a high interest-rate regime accompanied by escalating geopolitical and climate shocks, the incentive for automation is obvious even if the investment climate constraints the capacity to immediately act on it. If labour costs are rising due to labour shortages and (hopefully) successful trade union militancy then this further incentivises automation in the interests of reducing dependence on labour. This is ChatGPT’s useful summary of how high-interest rates discourage capital investment in automation:
- Higher Borrowing Costs: In a high-interest rate environment, the cost of borrowing money increases. Businesses looking to invest in automotive technologies typically require substantial capital investments, such as purchasing new machinery, developing infrastructure, or conducting research and development. Higher interest rates can make borrowing more expensive, increasing the cost of financing these investments. As a result, businesses may be less inclined to take on additional debt or postpone investment decisions due to the higher cost of capital.
- Increased Discount Rates: In finance, discount rates are used to evaluate the present value of future cash flows. A higher interest rate environment typically leads to higher discount rates. When businesses analyze the potential returns of their investments in automotive technologies, they discount future cash flows at higher rates. This means that the value of future cost savings and productivity gains may be reduced in present value terms, making the investment less attractive.
- Uncertainty and Risk Aversion: High-interest rate environments are often associated with economic uncertainty and market volatility. Uncertainty can make businesses more risk-averse and cautious about making long-term investments, particularly in capital-intensive sectors like automotive technologies. Uncertainty about future interest rates, economic conditions, and consumer demand can create a reluctance to commit significant capital to projects that have long payback periods.
- Opportunity Cost: In a high-interest rate environment, the cost of capital increases not only for businesses but also for potential investors. Investors have alternative options for deploying their capital, such as fixed-income investments, bonds, or other financial instruments that provide attractive returns with lower risk. As the cost of capital rises, businesses developing automotive technologies may face stiffer competition for investment dollars. Investors may allocate their funds to alternative investments that offer better risk-adjusted returns, leading to a reduced pool of available capital for automotive technology projects.
- Payback Period and Return on Investment: Automotive technology investments often involve a long payback period, particularly for significant infrastructure or research and development projects. In a high-interest rate environment, the extended time required to recoup the investment may lead to a lower return on investment when discounted at higher interest rates. This can make the investment less financially viable or require significantly higher expected returns to justify the risk.
But how much capital investment is needed for generative AI? My concern is that the most hucksterish elements of the platform economy will thrive under these conditions, offering out of the box solutions which purport to automate business processes. Rather than high interests rates prohibiting automation, I worry they might instead lead to automation on the cheap, with utterly chaotic intra-organisational consequences.
Mark Carrigan 