Artificial intelligence (AI) fueling 4.0 Industrial Revolution has put on center stage several questions in the context of a converging economy, such as the one of Slovakia. Will introduction of AI help the main pillars of the Slovak economy such as FDI-owned manufacturing industries ? Which policy challenges will it bring to the fore ?
Economic convergence leads to rising wages
Convergence of a national economy to the level of Western economies in the European Union means that via labor productivity increases overall economic growth surpasses that in the EU – overtime the gap between the level of economic performance of Western and Eastern countries decreases. With an overall national wage growth – as some key sectors over-perform – some sectors may run out of potential for productivity gains and its labor productivity tempo might be lagging behind the national average of wage growth.
Which are the two scenarios for development of such sector i.e. sector where labor productivity growth no longer goes hand in hand with wage growth at the national level ?
A) Scenario without much introduction of new machinery
Assume labor compensation/output remains at significant levels (say 40-50%) as is oftentimes the case in manufacturing sectors. Given that wages are rising at the national level, the company X must somehow upgrade the production so that supposedly positive gap between return and cost of capital is not eaten up. This can happen either through new processes or introduction of new higher-value products (such as higher value added car models in a car industry). This continuous upgrade of production must go on until the labor productivity growth reaches its limits. At that point, the factory could continue producing but the economic rate of return (ROIC- cost of capital) would start shrinking. Or it can re-allocate production to some other place such as further East. Given that labor constitutes a significant share of output, much lower labor costs in the East in excess of the re-allocation costs may drive decision to move eastward, to a cheaper location.
B) Scenario with a heavy introduction of new machines, incl AI
With introduction of labor saving new machinery such as AI machines, labor compensation to output ratio would likely plummet (let’s say from 40-50% levels to, say, 10% level) – hence cost advantage of labor diminishes since now labor accounts of much smaller share in cost structure. Whether it is Slovakia or low wage Kazakhstan, there is not much difference in labor cost to produce. However, the costs of re-allocation to the East remain still significant.
On the basis of this simplified analysis, it is clearly in the interest of Slovakia that foreign owned car plants introduce labor saving machines since this reduces the incentives that headquarters re-allocate production to cheaper locations, such as in the East.
This is because after AI machines are employed, wage bill/output ratio on a company level declines dramatically to the point where it nearly does not matter where the factory is located geographically. If now labor accounts for let’s say 10% of produced economic value this means that whether the factory is in France or Slovakia or Kazakhstan matters much less than before from the viewpoint of labor costs. However, given that it is costly to set up a new factory in the East, it is not very likely the car factory would migrate eastward (Kazakhstan). If anything, given the advantages of being in home country it may rather go back to home country (France) despite higher hourly labor costs.
In order to summarize, on the basis of this simplified analysis it seems that sectors where introduction of AI factories is feasible and makes economic sense, their installation would reduce incentives to migrate eastward. On the other hand, a significant gap between domestic wage level and a wage level in the East may create a very sharp incentives to re-allocate whenever the labor costs constitute major share of costs, as is the case in sectors without AI machines introduction.
The good news and the bad news for Slovakia
At a nutshell, for Slovakia AI introduction will be a good and bad news at the same time. It will be good news because AI will lead to reduction of importance of labor costs in certain sectors as the competitive factor and will blunt incentives for the existing manufacturing plants to leave Slovakia and move eastward where wages are lower than in Slovakia. In such case, geographical location is much less important and given the existence of re-allocation costs incentives will likely be to remain intact geographically.
However, it also means that for a country such as Slovakia future FDI will be much harder to acquire since the advantage of low labor cost as a competitive factor will be significantly reduced in AI intensive sectors. Things like quality of business environment, tax regime or reduction of country risk ( which will reduce the cost of capital) through better institutional environment will matter relatively more. More generally, it means we might thus move into the world where soft factors such business/institutional environment factors matter more for investment decisions than the labor cost.
Sector-level wise, at the national policy level in Slovakia one should focus efforts to attract FDI on areas where AI will not make much difference and a ratio of wage bill to output remains significant (such as tourism, hospitality, some other services). This is because attempts to lure FDI – in absence of strong advantages in the institutional arena – where wages do not constitute much share of output anymore would likely be futile.
AI likely to suppress incentives for FDI
More generally, application of AI to manufacturing will likely lead to reduction of downhill FDIs globally in the future and probably mean a less integrated world going forward. Pending AI machines installation projects may even already explain some of the recent fall in cross-border FDI in 2018. The world with much less FDI puts a premium on domestic economic policies to nurture home companies and fuel domestic economic growth.
The quality of human capital will continue to matter and its relative low cost may be a driver for certain inward investments (such as shared services) in sectors where there is still a high share of labor compensation on output. Investment in country’s human capital should thus remain important – after-all, AI machines dominated factories will require presence of super sophisticated managers/engineers. Tax incentives for support of educational sector would go a long way towards further beefing up domestic educational sector which should prepare such experts. Given that future incoming FDI will likely slow down – at least in those sectors where robots can replace humans – much higher effort should be focused on building a domestic enterprise sector via start-ups support and different schemes of nurturing domestic entrepreneurship.
A divergence back again ?
If the simplified analysis above is correct, implications of it seem quite dim for developing countries and challenges they are facing. Given that their business and institutional environments are relatively weak and labor cost is a primary competitive advantage factor, FDI where robots can replace humans will likely be not lured to such areas to the same extent as before. Some FDI factories might even migrate uphill to areas with better business, tax and institutional environment since importance of labor cost will be so drastically reduced. In developing countries, bottlenecks on markets with super-sophisticated managers/engineers to run such AI-intensive factories may contribute to that as well. Weak business and institutional environment, underdeveloped human capital all leading to now much lower intake of FDI and a low sophistication of domestic economies to begin with, mean that such developing countries might face challenges to develop modern export sectors and thus to follow development model based on export-led growth focusing on sophisticated sectors. It could also lead to the race to the bottom regarding environmental standards as developing countries try to compete on non-labor cost factors. Whether AI heavily applied to manufacturing sector also means a higher probability of middle-income trap incidence for countries such as Slovakia is worth pondering too.
Vladimir Zlacky, LookingEast.eu
Bratislava, 26 June 2019