Economics Catalysts

It is often said that ‘money makes the world go around’, although we need to examine this adage in a little more detail. For if money is little more than an IOU backed by the perceived wealth of some nation-state, then we might need to ask:

By what ‘yardstick’ is the wealth of a nation measured?

We might initially make some assumptions about its access to natural resources, the skills and education of its population and the quality of its economic infrastructure, inclusive of both its industrial and service sectors. We might then attempt to abstract these initial assumptions in terms of natural, human and infrastructure assets on the nation’s balance sheets, which are then offset by any liabilities in terms of public and private debt. Some economists may then judge the level of debt in terms of its ratio to the Gross Domestic Product (GDP), although GDP is more of a measure of current national income rather than necessarily being an accurate measure of future wealth. Today, we might recognize that the wealth of some nation-states is primarily based on natural resources in the form of fossil fuels, which are perceived as a commodity to be sold on the global market and subject to fluctuating supply and demand criteria. However, we might table a question at this point:

What if these natural resources were made worthless by technology innovation?

We might reasonably assume that without some alternative source of income, the on-going wealth of these nation-states may be short-lived unless the deficit in the economy can be made up by its industrial or service sectors. Of course, even these sectors may be susceptible to future technology change, such that it is often assumed that the biggest asset of a nation-state is its population, at least, if well educated.

Note: The UN calculates what it calls ‘ human capital ’ based on a number of criteria, e.g. population, education, employment prospects and years to retirement. For example, human capital has been calculated to represent 88% of the UK’s wealth and 75% of US wealth.

While this opening discussion is not attempting to be too exact on the actual complexity involved, we might start to see an argument that the population might actually be the biggest wealth asset in a given economy. This said, it might also be realised that future AI and robotic automation could dramatically change the wealth value associated with human capital in a similar fashion by which the value of natural resources might be revised based on supply and demand. However, in the human context, the issue of supply and demand may be better described in terms of the quantity and quality of a given population, which starts to stray into a discussion that political correctness might prefer we simply ignore. However, if the truth be told, it is unclear that man-made change will be any kinder to specific individuals than natural selection when survival is at stake, such that the issues implied in the following questions may still be a factor that shapes the future.

What percentage of the population is intelligent, educated and productive?
How might this change in the future in respect to AI and robotic developments?

Taken together, we might realise that the second question places an inference on how we might answer the first question, at least in terms of the productive aspect of a given population. However, there is also an inference that only a percentage of any given population will maximise education, if defined in terms of its IQ distribution. While many will refute the idea that the value of humanity can be quantified in terms of a one-dimensional IQ score, the changing demands of employment suggests that it cannot be ignored. As an initial and overly simplistic summary, an IQ score between 90-110 might be considered as an average range attributed to 50% of the population, where approximately 25% have a higher score and the remaining 25% have a lower score. However, there is already substantial evidence that intelligence is an important indicator of the type of employment that can be realised. In this context, IQ reflects an ability to deal with the cognitive complexity associated with different types of productive employment, which is generalised in the following chart.

However, this chart is essentially historical in nature in that it does not account for the potential for AI and robotic automation to increase across all employment sectors in the coming decades. If so, there is a suggestion that such technology developments will impact a much wider range of blue and white-collar professions in the future.

Note: If AI automation continues to progress, even at present rates, do we need to consider the probability that unemployment will increase over time? If so, this may lead us to other far more uncomfortable questions, e.g. how will the growing population of unemployed survive, not just economically, but socially and possibly even physically? While it is accepted that the implications of this last question may, yet again, sound the alarm bells in a politically-correct world, it is unclear that simply ignoring the issue will provide any sort of solution.

In the following table, a selection of occupations in the US job market are qualified with a risk probability of job losses against the ‘current’ employment figures in millions. As a comparative benchmark, the number of working age (16-64) people in the US is estimated to be in the region of 210 million of which only about 130 million (62%) are estimated to have full-time employment. While the official unemployment rate in the US is stated to be less than 5%, it is doubtful that this figure truly reflects the scale of mental and physical disabilities that prevents many gaining meaningful employment, which is then compounded by part-time employment and zero-hour contract work. Therefore, if we simply use the 62% figure, this would mean that something like 80 million (38%) of the US working population is not in full-time employment, while the job losses shown in the following table suggest that the US may require an additional 30 million new jobs by 2050 to offset losses, even if population growth is ignored.

  Occupation Current
Employment
Risk
Probability
Job
Losses
1 Retail Sales 4.3 92% 3.96
2 Cashiers 3.3 97% 3.20
3 Fast Food Services 2.9 92% 2.67
4 Teaching 2.9 56% 1.62
5 Office Clerks 2.8 96% 2.69
6 Waiters & Waitresses 2.3 94% 2.16
7 Customer Services 2.3 55% 1.27
8 Nursing 2.2 1% 0.02
9 General Labour 2.1 85% 1.79
10 Janitors 2.1 66% 1.39
11 Administration 2.1 96% 2.02
12 General Managers 1.9 16% 0.30
13 Stock Control 1.8 64% 1.15
14 General Accounting 1.6 98% 1.57
15 Sales Representatives 1.4 85% 1.19
16 Maintenance & Repair 1.2 64% 0.77
17 General Supervisors 1.2 28% 0.34
18 Accountants & Auditors 1.1 94% 1.03
19 Security Guards 1.0 84% 0.84
20 Assembly Workers 1.0 97% 0.97
21 Restaurant Cooks 1.0 96% 0.96
  Totals & Average 42.5 74% 31.90

This table simply highlights some specific categories at risk, but not all. For example, in some urban areas up to 17% of the US working population are employed, both full-time and part-time, in driving vehicles of some description, e.g. buses, taxis, deliveries, haulage, which might all be at increased risk given the potential of driverless vehicles over the coming decades. There is also little recognition of the impact that cognitive AI systems might yet come to have on professional careers, such as lawyers, accountants and doctors plus many others. If this wider spectrum of white-collar professions is also at risk, it might be suggested that 25-50% of all current jobs may be at risk by 2050.

So, what would be the implications of such levels of unemployment?

Clearly, increasing unemployment may cause instability in all facets of the human ecosystem, i.e. social, political and economic, although for the moment this discussion is primarily focussed on the economic scope. Based on the previous description, we might realise that AI automation may increase the risk probability of job losses within white-collar sectors in a similar fashion that robotic automation has already caused job losses in the blue-collar assembly sector. As such, IQ and education may not necessarily be a guarantee of job security in the future, although it may still reflect a higher ability to more quickly retrain to meet the changing and competitive demands for new jobs. However, we might also realise that in this ‘brave new world’ people are not only competing with increasing AI and robotic automation, but also with an increasing number of people seeking productive employment, where supply and demand may be heavily biased towards the employer, not the employee.

But how would a future of increasing unemployment affect the economy?

Obviously, unemployment may have an immediate impact on an individual in terms of lost income, which may then spill over into the local economy in terms of reduced spending to businesses operating in the area. While some of these individuals might be able to offset the loss of income by accessing savings and/or applying for loans, increasing debt cannot be a viable solution for the individual or the economy.

Note: While the idea of gross domestic product (GDP) is not necessarily a definitive measure of the long-term health of an economy, it might still be used as a general indicator. We might define GDP=C+I+G+(X-M), where [C] is consumption of goods, [I] is private investment, [G] is government spending, [X] is exports and [M] equates to imports. As a general measure, we might assume that GDP also reflects the government income by way of taxation, which in a perfect world would cover all government spending without running a deficit leading to an accumulating national debt.

Today, the health of an economy is often quantified in terms of the percentage growth in GDP, which is a relatively complex interaction of the components defined above. We might readily see that increasing unemployment might affect the consumption of goods and require increased government spending in terms of the social welfare associated with unemployment. However, increasing unemployment may also trigger a long-term economic depression causing a fall in the value of the fiat currency on foreign exchange markets, which then increases import costs and a potential loss of confidence in private investment in general. 

Note: Okun's Law reflects an observed rule-of-thumb relationship between unemployment and GDP. This law suggests a 1% change in unemployment rate will correspond to a 2% change in GDP, both rising or falling. Without any attempt to justify this inference, we might simply table a question as to the economic implications of unemployment ever reaching 25-50% of the working population?

Despite the suggestion of dire economic consequences to both individuals and governments, many businesses may not be able to resist the necessity of further competitive gains that AI and robotic automation may offer to their bottom-line profits, at least, in the short-term. For, unlike people, AI and robotic automation will only represent an investment in capital cost, which can in part be offset by the tax-relief on depreciation, while reducing demands for wages, maternity leave and pension funds plus all the other HR overheads. If you think that this bottom-line focus is too cold and calculating, you might wish to consider how many of today’s multinational corporations attempt to ‘avoid’ paying the full rate of taxes that they know are required by governments to maintain social services. Of course, like technology, even multinational corporations may not necessarily be allowed to operate in complete isolation of mounting political and social pressures, such that we still need to provide an introduction of the nature and scope of such issues.