The shift to a more services-based economy increases economy-wide depreciation rates, and so the need to invest. Higher average interest rates – and economy-wide profit shares – are among the likely results.
An important reason for our house view that the global structure of interest rates will be higher going forward than it was pre-pandemic relates to the balance between saving and investment. A range of forces are pointing in the direction of higher investment, without an obvious counterbalance to boost saving at the same time. Among these forces are the energy transition and energy-intensive new technologies, including AI with its high compute requirements.
In a note earlier this week, Westpac Economics Senior Economist Pat Bustamante highlighted that some of that shift to higher investment is already evident in the Australian data. The industries that are most involved in the energy transition and the adoption of leading technologies are already increasing new investment, especially in software and other so-called ‘intangibles’.
As Pat also observed, there are two implications of these transitions that are not immediately obvious. First, the shift to a more services-based economy, away from traditionally capital-intensive industries such as manufacturing, does not necessarily mean that business investment is lower. Second, and relatedly, new investment is increasingly in the types of capital with higher rates of depreciation and obsolescence than traditional physical plant and machinery. Businesses must ‘run harder to stay in place’, lest their capital stock starts to diminish. Industries that are not obviously capital-intensive nonetheless may need to invest intensively. Pat’s note shows that, as an economy, Australia’s depreciation rate is rising, and has been for some decades.
To the extent that new investment adds to the capital stock and improves its quality, we can expect some payoff in the form of higher productivity and output growth. But the investment that is replacing existing capital is simply covering depreciation. While some of the new technologies may fall into the first category of productivity-enhancers, much of the investment into the energy transition is pure replacement of existing capital stock – effectively an accelerated depreciation. In this respect, there is a bigger payoff to investments that make other activities more energy efficient than to those simply replacing existing generation and distribution infrastructure.
Investing to replace depreciated capital or execute the energy transition is still worth doing. The costs of not doing so are large. But if the economy-wide depreciation rate on the capital stock has risen, this has other implications that are perhaps not widely understood.
To the investors go the spoils
If a higher depreciation rate partly occurs because of higher rates of technical obsolescence – as you’d expect with increased usage of software-based innovation, for example – then new investment introduces different types of capital. New skills may be expected of the workers using the newly-installed capital. More generally, if the optimal mix of labour skills and capital changes as new capital replaces old, then a faster rate of technical change and obsolescence means a faster rate of churn in the kinds of jobs available.
We saw the same thing happen in the first wave of the software revolution. The adoption of PCs and later the internet accelerated obsolescence rates, as did the increased integration of software elements into traditional physical capital. The result was increased physical churn in the capital stock, but also in the skills needed of workers. This lowered the bargaining power of workers and shifted some of the share of income from production away from wages towards profits, especially in countries where there were also barriers to entry for new firms.
Or at least, this is one of the possible explanations of the upward trend in the profit share (downward trend in the wage share) seen in a range of industrialised economies from about the mid-1980s to just before the Global Financial Crisis. And nearly two decades after proposing that explanation in a paper I wrote with former RBA colleague Kathryn Smith (partly based on prior work by Hornstein, Krusell and Violante, subsequently published here), it’s still the explanation that I think makes most sense. To be fair, there are other hypotheses that also fit some aspects of the data, but the hypothesis in that paper explains the timing and cross-country pattern in the trends, in a way that some other explanations do not.
In particular, the nexus between capital obsolescence rates, labour market churn and income shares helps make sense of the end of that upward trend in the profit share in the mid-2000s. Across advanced economies, the post-GFC period was one of low private investment, low productivity growth – and little apparent trend in the profit and wage shares. In Australia, for example,
RBA analysis shows that the profit share outside mining has been broadly flat for two decades. This fits in with the idea that the earlier upward trend in the profit share was at least partly explained by the wave of adoption of an earlier generation of IT products, and that by the mid 2000s, this wave had matured.
If we are indeed on the cusp of a period of faster replacement of existing capital, and some of that demands new skills of workers, it’s possible that we will see this trend increase in the profit share (decline in the wage share) resume. That might be good for productivity growth, but it is not guaranteed that real wages growth will keep pace.
At the least, it is a reason to be cautious about wages forecasts and avoid being too bullish. This is especially so in a country like Australia, where wages growth undershot official forecasts for years, even with a flat trend in the share of wages in national income.
