The productivity challenge in an era of weaker foreign labour growth
By Michael Wan, Analyst @ Credit Suisse
The growth in Singapore’s population and its dependence on foreign labour have become a major issue in the country. In this report, we analyse what impact the government’s restrictions on foreign labour might have on trend GDP growth.
What is the impact on trend GDP growth?
The growth in Singapore’s population and its dependence on foreign labour have become a major issue in the Island state. It is certainly a focus for the government, which has fleshed out its latest thinking in two recent reports.
First, the release of the Population White Paper, which detailed the government’s longterm plans on the topic. The headline population projection of 6.9 million by 2030, from 5.3 million currently, which the ruling party termed its ‘worst-case scenario’, garnered strong and emotive responses on the potential impact such a “large” population would have on social well-being. Ironically, we have noted that the projections, if they are met, would actually require a sizeable slowdown in the growth of the foreign labour force.
Second, the 2013 Budget intensified the push to restructure the economy by introducing more foreign worker tightening measures, including higher salary and educational requirements, additional foreign worker levies to sharpen the distinction between skilled and unskilled workers, and a lower ratio of foreign to local workers a company can hire.
Both these events are indicative of the government’s attempt to shift the country away from its traditional reliance on population growth to fuel economic expansion. The growth in the total labour force is officially expected to slow significantly over the medium term from current rates of 3 per cent to around 1.5 per cent per annum, due to the combined effects of additional foreign labour curbs and an ageing population. With this in mind, we tackle two questions of vital significance to Singapore’s long term economic future in this report:
What impact will the foreign labour restrictions have on trend GDP growth?
The limits on foreign labour have been stepped up since 2010, while there are a slew of additional restrictions due to be implemented from 2013 to 2015 (see Exhibit 4 for details). To the extent they work, the restrictions effectively represent a negative supply shock with potentially important implications for the growth of the productive potential of the country.
So how big is the impact on trend GDP growth likely to be? We note that to meet the government’s population projections in the White Paper, the growth in the foreign labour force would have to slow to about 30,000-40,000 workers per annum on average over the next 8 years from current rate of around 80,000. We have incorporated this reduction into our projections of labour force growth over the next 8 years, which, if these curbs succeed, would ultimately translate into around 280,000 fewer foreign workers than would otherwise have been the case by the end of the period (see Exhibit 5).
In order to gauge the overall reduction in the supply of labour, we really need to project the potential reduction in total person-hours worked (not just employment) due to these foreign labour curbs. To do so, we have incorporated several other assumptions in our projections, namely, 1) the share of labour across various sectors (manufacturing, construction and services) remains constant till 2020, and 2) the average number of hours worked across various sectors remains constant till 2020. On the basis of these assumptions, we estimate that the number of person-hours worked would be 7 per cent lower by 2020 compared with a base case where the foreign labour force continues to grow at its previous rate (see Exhibit 6).
On this basis, we estimate that the foreign labour force restrictions could shave 0.9 percentage points off Singapore’s annual trend GDP growth (see Exhibit 7).
This reduction is fairly sizeable, but does assume labour productivity growth and the labour force participation rate remain unchanged. If either or both move significantly higher, then the direct effect of the restrictions could be countered. So what are the chances of this happening? We look at the prospects for productivity growth and the participation rate in turn.
What could potentially offset the fall in trend growth?
#1 Stronger labour productivity growth
To what extent could an increase in labour force productivity growth offset the decline in labour force growth over the next few years?
While in 2010, productivity growth surged, it has subsequently slowed dramatically (see Exhibit 8), and we note that the level of labour productivity in the overall economy is no higher than it was back in 2005 (see Exhibit 9). This is turn could suggest that trend productivity growth has actually slowed – not exactly what the government was hoping to see.
Looking within sectors, it is clear that manufacturing, as one would expect, has been the key sector driving labour productivity growth over the past two decades. While manufacturing productivity fell sharply during the Global Financial Crisis, its level has since returned to its previous trend (see Exhibit 10). Although construction sector productivity has been on a general downward trend, it has been rising gradually over the last 7 years (see Exhibit 11). By implication therefore, the key reason for the weak performance of labour productivity over the past 7 years has been the services sector, with productivity there currently 8 per cent below its long-term trend (Exhibit 10).
With this weakness in services productivity in mind, we believe the key question to ask is whether the recent below-trend growth in services sector labour productivity is purely cyclical, or reflects a structural deterioration? To help answer this question, we have used a technique called shift-share analysis.
Some answers using shift share analysis
Shift-share analysis allows us to decompose changes in productivity into the following factors:
1) The within effect (intra-sectoral): The contribution to productivity growth in individual sectors after adjusting for inter-sectoral changes. A rise in the “within-effect” could potentially reflect capital deepening or more efficient adoption of technology within a sector.
2) The static shift effect (inter-sectoral): This will be negative if there is an increase in employment share to sectors with comparatively lower levels of productivity, and vice versa. An example of a negative “static shift” would be a rise in employment share in the accommodation and food sector, which historically has lower levels of productivity.
3) The dynamic shift effect (inter-sectoral): This will be negative if there is a fall in the employment share towards sectors with comparatively higher labour productivity growth, and vice versa. For instance, we would see a negative “dynamic-shift” effect if there is a shift in employment share away from wholesale and retail trade, which has enjoyed a CAGR in productivity of around 0.7 per cent over the past seven years, compared with negative growth for most other services sectors over the same period.
There has been a shift towards services sectors with lower levels of labour productivity
As Exhibit 11 shows, the ‘within effect’ (intra-sectoral) has historically been the main driver of labour productivity growth in services sectors. However, this effect appears to capture cyclical elements, for instance, the bounce in productivity post the Global Financial Crisis (see Exhibit 12).
More importantly, while the ‘static shift’ effect (inter-sectoral) has historically been positive, it has turned negative from 2007, contributing to the slowdown in labour productivity growth (see Exhibit 13). In other words, there has been a distinct shift towards services sectors with lower levels of labour productivity.
Looking more closely within the services sector, and with the exception of the finance industry, the increase in labour shares in recent years has generally come in services sectors with lower levels of labour productivity, such as the business services sector (e.g., real estate), the accommodation and food sector (hotels and F&B), and the other services sector (e.g., health care and administrative). These are mainly the domestically oriented sectors (see Exhibit 14 below). In comparison, the externally oriented services sectors, which generally have higher levels of productivity (see Exhibit 15), have generally reduced their employment shares since the financial crisis.
In our view, this shift in focus towards domestically oriented services sectors is here to stay, reflecting not just the weakness of the export sector since the Global Financial Crisis, but also the impact of policy, as the government has actively encouraged the development of more “domestically” oriented sectors.
One of the key reasons why productivity levels tend to be lower in domestically oriented services sectors than in externally oriented ones (see Exhibit 15) could be that it is more difficult to substitute capital for labour in the former compared to the latter. According to an MAS study, the ease with which capital can be substituted for labour is the highest in externally oriented services sectors such as wholesale and retail trade (95 per cent), compared with domestically oriented ones such as the hotel and restaurant sectors (21 per cent)3 (see Exhibit 16). This would in turn help explain why, in the process of restructuring, the services sector has not been able to reduce its dependence on foreign labour force growth as fast as the export sector (see Exhibit 17).
As such, with the increasing focus on sectors which happen to have lower levels of labour productivity, it seems to us that trend productivity growth in the services sector has indeed fallen and will be constrained moving forward.
Manufacturing sector productivity appears to be fine, but there are a couple of concerns
While, as we have noted, productivity growth in the manufacturing sector appears to be fine, we note a couple of worrying trends that could imply an eventual shrinkage in the size of the manufacturing sector relative to the economy as a whole. This would be further bad news for trend productivity growth in the economy, given that productivity levels are already far higher in manufacturing compared with most other services sectors, coupled with the fact that it is relatively much easier to raise productivity in manufacturing given the ease of substituting capital for labour (see Exhibit 16).
While these two issues could reduce the importance of the manufacturing sector within the economy and, hence, reduce trend productivity growth, we note that Singapore has had a very good record of diversifying its manufacturing base and encouraging new industries to develop. It is entirely possible that another sector, even beyond pharmaceuticals and offshore and marine engineering, will eventually pick up the slack.
Comparison with other developed countries indicates that Singapore’s productivity drive could be a slow process
While it is true that the level of productivity in Singapore is some way below that of the global leaders such as the United States (see Exhibit 20), the experience of other developed countries shows that raising productivity levels is often a slow process. We note that most developed countries, which already enjoy a high level of productivity, do not usually attain the 2 per cent-3 per cent productivity growth rates that the Singapore government targeted in its White Paper. For instance, based on data from the US Conference Board, while France, Denmark and Finland have a similar level of labour productivity as Singapore, these countries have only managed to achieve around 0.2 per cent-0.3 per cent annual growth in labour productivity since 2005 (see Exhibit 21).
There are generally two reasons why developed countries, such as Singapore, tend to have fairly low trend productivity growth. First, given the already high capital stock in rich economies, it is difficult to achieve the incremental returns on capital that one might expect to see in a less developed country with a lower capital stock. Put simply, there is less need to build infrastructure, for example. Second, we note that developed countries tend to have an aging population, which may arguably result in lower trend labour productivity growth.4 In addition, we note that Singapore already ranks as one of the countries with the highest average number of working hours per week at 44 hours, with only Hong Kong coming in higher at 45 hours, according to data from the US Conference Board. In comparison, people in South Korea and Taiwan work 42 and 41 hours per week respectively, on average. This in turn suggests that Singapore will find it difficult to increase output per worker by increasing the time people spend at work.
Productivity gains will probably remain elusive over the next few years
In conclusion, while productivity growth in Singapore is likely to register a cyclical improvement as and when the global economy shows a meaningful pick-up, productivity growth looks set to remain constrained over the next few years, given the inter-sectoral shifts towards industries with lower levels of productivity. It might take years before we are able to see substantial improvements in trend productivity growth in these areas, given that, as we pointed out, it is much more difficult to substitute capital for labour here. In our view, 1 per cent per annum trend productivity growth is a more realistic projection than the government’s 2 per cent-3 per cent objective for the rest of this decade.
What could potentially offset the fall in trend growth?
#2: Labour force participation rate
With trend labour productivity growth unlikely to offset the potential shock to Singapore’s labour supply, the next question is whether a rise in the resident labour force participation rate, defined as the percentage of economically active persons to the total population aged fifteen and above, could offset the reduction in the foreign workforce.
We have investigated the implications of two scenarios here: 1) the labour force participation rate rises from the current 67 per cent to 74 per cent − the rate for Iceland which has the highest participation rate among high income countries according to the World Bank − and 2) the participation rate increases to 70 per cent (the rate for Switzerland).
We note that even our assumption of a 74 per cent labour force participation rate would not generate quite enough extra workers to offset the projected reduction in the foreign workforce as targeted in the White Paper (see Exhibit 22). A rise in the participation rate to those levels would imply about 30,000 additional workers in the labour force every year, compared with the projected 35,000 reduction in foreign worker inflows each year (see Exhibit 22 below). Meanwhile a 70 per cent labour force participation rate would imply a much more subdued 15,000 additional indigenous workers per year.
Even if the labour force participation rate were to rise substantially, we highlight that there are likely to be significant mismatches between the expectations and skills of the local workforce compared with those of foreign workers. In particular, out of the job vacancies last year, companies found it much more difficult to attract locals to fill non-PMET (“lowskilled”) jobs, such as cleaners, labourers, and plant machine operators, compared with the PMET (“high-skilled”) ones (see Exhibit 23). In addition, while it might be possible to increase the labour force participation rate by attracting people currently in school or retired, or by raising the female labour participation rate, many of these people typically prefer to work part-time (see Exhibits 24 and 25).
Ultimately, it seems to us that Singapore cannot depend solely on a rise in the labour force participation rate to counter the reduction in foreign workers.
Putting it all together – The implications for trend GDP growth
We expect medium-term trend GDP growth to be around 2.5 per cent
Given our view that labour productivity growth will remain structurally weak (around 1 per cent) for some time, and the difficulty involved in raising the labour force participation rate as well as the average length of the working week, we believe the government’s restrictions on the growth in the foreign labour force will mean trend GDP growth is lower over the next few years.
With a medium-term labour force growth assumption of 1.5 per cent, down from about 3.5 per cent over the last decade, and a labour productivity growth assumption of 1 per cent, compared with 1.4 per cent over the past decade, we estimate that trend GDP growth will be around 2.5 per cent over the medium term, nearly half that of the previous decade (see Exhibit 26 below).
If we are right, Singapore is going to have to get used to a much lower rate of economic growth in coming years, from what was a very impressive trend growth rate for a developed country. In a sense, Singapore may finally be becoming “normal”.
We highlight a few key implications stemming from this conclusion:
1) Worse growth-inflation trade-off. For a given level of GDP growth, Singapore is likely to experience higher inflation compared to the past. For example, while inflation has moderated from the highs seen at the start of last year, any cyclical pick-up in growth later this year, perhaps due to an improvement in the global trade cycle, could add to inflationary pressures.
2) Lower FDI. While for decades Singapore has seen very strong foreign direct investment (FDI) inflows compared to its peers in the region, the government is likely to be more selective in the types of FDI it wants to attract moving forward, given the restrictions on the growth in foreign labour and increasing land constraints. Indeed, the Economic Development Board (EDB), the main government agency designated to attract investment into the country, is targeting a lower level of inward FDI in 2013 from the previous year.
3) More moderate property price rises? Given much lower trend GDP growth, property, which has been an important asset class to many Singaporeans, could see more moderate capital gains. This of course will depend on how the supply of housing evolves but, interestingly, Singapore’s minister in charge of housing suggested property price gains are indeed likely to be lower in the years ahead.
4) Constraints on fiscal revenue and spending. Slower trend growth would constrain fiscal revenue and, in turn, spending, if the government maintains its traditional budget surplus.