The Census and Statistics Department is responsible for compiling projections of the population of Hong Kong. The main use of the projections is to provide a common basis for Government programme planning in such areas as education, housing, transport, social services and healthcare services. Projection results are also widely applied in academic research, and business development and planning in the private sector.
A set of population projections, with mid-2014 population estimates as the base and covering the period 2015-2064, has been prepared and released on 25 September 2015.
Based on the results of the new set of Population Projections, the Census and Statistics Department has also updated the labour force projections and domestic household projections.
Preamble
The Census and Statistics Department (C&SD) released the latest set of population and labour force projections in September 2015. The projections form a common basis for the Government in planning various public services such as housing, education and healthcare services. In compiling population and labour force projections, C&SD has adhered to professionalism and adopted an objective and impartial attitude. The methods adopted were similar to those used by other advanced economies, and the assumptions were solely based on professional judgments. Since a lot of technical details were involved in population and labour force projections, we would like to clarify some common queries.
Was the timing of releasing the projection results arranged to facilitate certain policy recommendations or consultations, such as the controversial retirement protection consultation?
According to the practice in the past 20 years, C&SD has updated population and labour force projections every 2 to 3 years. Results of the latest population projections were released in September 2015 while those of the previous round were released 3 years before, i.e. in July 2012. Prior to that, population projection results were released in 2010, 2007, 2004 and 2002 respectively. As in the usual practice, population and labour force projections compiled by C&SD provide a common basis for the Government in planning various public services such as housing, education and healthcare services and are widely used by different Government bureaux/ departments. The projections also serve as impartial reference information for people from various sectors of society who have analytical needs. C&SD will not conduct population and labour force projections to facilitate deliberation of a single policy. Regarding the public consultation on retirement protection launched by the Government in December 2015, population and labour force projection results were only one of the many data inputs used in the estimation of the financial burden.
Was the projection period extended from 30 years to 50 years to exaggerate the financial burden of retirement protection?
Extension of the projection period was solely based on professional considerations without any political influence. In this round of population projections, we originally intended to continue using the 30-year projection period. However, population was projected to decline as from 2043 (i.e. 28 years later). In order to help the society discern whether the decline was temporary or persistent, there was a need to extend the projection period. What should be the appropriate span for the projections? Making reference to the practice of other places, it was observed that many advanced economies were adopting projection periods of around 50 years, including Japan, Korea, Singapore, Taiwan, the USA and Canada. As such, we extended the projection period from 30 years to 50 years.
Compared with the last round, the latest projections have only incorporated additional data of the recent 2 to 3 years. Why was the labour force substantially reduced in the next 50 years?
Compared with the labour force projections released in September 2013, the projected labour force (excluding foreign domestic helpers) in the current round was lower than that in the last round by 3% (i.e. some 0.11 million persons) at the end of last round’s projection period (i.e. 2041).
The downward revision was mainly due to the reduction of the number of One-way Permit holders assumed to be coming to Hong Kong, from 150 persons per day in the last round to 100 persons per day in the current round on a long-term basis (i.e. from mid-2027 onwards). This assumption was formulated based on the actual inflow excluding overage children during the past 3 years. Given that the objective of the One-way Permit Scheme was family reunion, C&SD has studied the trend of Mainland-Hong Kong cross-boundary marriage during the past 30 years. Some structural changes were observed, including an increase in the proportion of Hong Kong females whose Mainland husbands tend to have lower probability of coming to Hong Kong (as compared to Mainland wives), and a probable decrease in the offspring from cross-boundary marriage. In view of these structural changes, the number of One-way Permit holders coming to Hong Kong is not likely to rebound to 150 per day. Moreover, overage children coming to Hong Kong is only a temporary phenomenon. In this connection, it was reasonable to adopt the long-term assumption of 100 One-way Permit holders coming to Hong Kong per day.
In the projections, was it true that only factors leading to decrease in labour force were considered while the effects of policies leading to increase in labour force were ignored?
Some critics have alleged that in conducting the projections, C&SD only considered factors leading to decrease in labour force, such as reduction in the number of One-way Permit holders coming to Hong Kong. On the other hand, the effects of policies leading to increase in labour force, especially those aimed at attracting talents and boosting fertility, were all assumed to be zero. Some people even suspected that C&SD has deliberately reduced the labour force projections so as to exaggerate the financial burden arising from retirement protection.
We have to point out that in conducting population and labour force projections, C&SD has considered all factors affecting the population trend, regardless of upward or downward effect and formulated all related assumptions based on objective data.
For policies which have been implemented for a period of time, their impact has already been reflected in the past trend and thus taken into account in the population and labour force projections. As for policies which were newly implemented or not yet implemented, since their impact had not yet materialised in the objective data available, they were not considered in this round of population and labour projections. C&SD will closely monitor the latest developments of the demographic and labour force situation and update the population and labour force projections as necessary.
(a) Upward adjustment of labour force participation rates
The projected labour force is basically derived by multiplying the projected labour force participation rates (LFPR) to the corresponding projected population. C&SD not only has not deliberately reduced the labour force, but has improved the projection methodology by considering the factor of educational attainment. This enhancement has led to an increase in the LFPR of most working age groups in tandem with the general rise in educational attainment. Overall, the projected LFPR in 2041 in the current round was 50.1%, which was higher than that in the previous round (49.5%) by 0.6 percentage point. Noticeably, the LFPR of females aged 50-54 and males aged 35-39 have increased by 3.4 and 2.0 percentage points respectively.
(b) Impact of policies should be based on objective data
Regarding the allegation on policy impact, we have to emphasise that in the past population and labour force projections, it has been a standing practice not to take into account impact of new policies which has not yet been reflected in actual data. This is because in the absence of objective data, subjective judgment has to be used. This will not only affect the credibility of official statistics, but is also against the “evidence-based” principle in the statistics profession.
Policies attracting talents
Some critics have doubted whether the inflow of labour force under existing policies was fully considered in the population projections. For instance, around 30 000 visas were granted under the “General Employment Policy” each year. Why was the related net inflow only several thousands? It should be noted that the mobility of talents from outside Hong Kong is very high. While a large number of new entrants are admitted into Hong Kong every year, those leaving are also significant in number. The average annual net inflow is thus in the magnitude of several thousands only. For example, in 2012 to 2014, the average net inflow of entrants under the “General Employment Policy”, “Admission Scheme for Mainland Talents and Professionals” and “Quality Migrant Admission Scheme” was about 3 800 per year. In the population projections, after taking into consideration the actual figures and past trends on the net flow under these 3 schemes and other schemes, the long-term net inflow of Hong Kong Non-permanent Residents other than foreign domestic helpers is projected to be around 4 600 persons per year. If there is no major change in the existing policies, a substantial increase in the net inflow figure in future is not likely, and the assumption is considered reasonable.
As for newly implemented policies such as the “Admission Scheme for the Second Generation of Chinese Hong Kong Permanent Residents”, they were not considered in this round of projections because of the lack of objective data on their impact at this moment. We will closely monitor developments in this regard and determine how to take into account the impact in the next updating of the population and labour force projections.
Policies boosting fertility
Some critics have also queried why the total fertility rate will not rebound substantially given the series of policies which aimed at fostering a supportive environment for forming and raising families recently launched by the Government.
It should be noted that women’s fertility trend in Hong Kong, influenced by its cultural background, is similar to that in other parts of Asia. The effect of policies boosting fertility is quite different from that in western countries. For instance, Singapore has implemented a series of policies to raise fertility, including direct cash allowances and funding programmes to help employers defray the implementation costs of family-friendly measures. However, significant rebound in fertility rate has not been observed and the total fertility rate has remained at the level of around 1 200. The total fertility rates of Korea and Taiwan have also been around 1 200 in recent years. In addition, Hong Kong’s total fertility rate has remained below 1 200 most of the time during the past 20 years (exceeding 1 200 in only 4 years out of the 20 years). Thus, a substantial rebound in Hong Kong’s total fertility rate in the future is not likely based on the recent trend.
Was the proportion of never-married women deliberately inflated from the current one-seventh to around 30% 50 years later so that the fertility rate and labour force would decrease?
There has been a rising trend of spinsterhood over the past 2 decades (see chart below). According to statistics in past Population Censuses/By-censuses, in 1991, only about 3% of women remained never married at age 49. The proportion rose to about 12% in 2011. Further analysed by educational attainment and age group, we have the following observations:
(1) During 1991-2011, the proportions of women remaining never married at age 49 increased for all levels of educational attainment. In 1991, the proportions were 1.7%, 4.1% and 5.8% for women with lower secondary and below, upper secondary and post-secondary educational attainment respectively. The corresponding proportions rose to 6.1%, 14.7% and 20.0% in 2011.
(2) Compared with women with lower educational attainment within the same age group, the never-married proportions of women with post-secondary educational attainment were not only higher, but exhibited the greatest cumulative increase during the period.
(3) Analysed by age group, there was an upward trend in the never-married proportions of women for all age groups during the past 20 years. This reflects that more and more women tend to postpone marriage or remain unmarried. Concurrently, the median age of first marriage increased from 26.2 in 1991 to 28.9 in 2011.
In fact, a rising trend in the proportion of never-married women is observed in other advanced economies. Regarding the proportion of never-married women aged 45-49, in Canada, the proportion increased from 7.4% in 1991 to 19.4% in 2014. In Australia, the proportion increased from 5.0% in 1991 to 14.4% in 2011. In Japan, the proportion increased from 6.3% in 2000 to 12.6% in 2010. The increasing trend was similar to that in Hong Kong (from 3.1% in 1991 to 13.6% in 2011). We can see that in tandem with the general rise in the level of educational attainment of women, increasingly more women are independent economically. Postponement of marriage and spinsterhood has become more common. The rising trend of the proportions of never-married women in Hong Kong is reflecting this phenomenon.
In Hong Kong, the proportion of never-married women at age 49 increased to 14% (around one-seventh) in 2014 and is projected to increase further to 31% (around 30%) in 2064. In other words, the actual proportion increased by 11 percentage points during the past 23 years (1991-2014) and was projected to increase by only 17 percentage points in the next 50 years. Given that the pace of increase in the next 50 years was already assumed to be slower than that in the past 23 years, there was no deliberate overestimation.
Were past labour force projections released by C&SD often inaccurate?
Some critics have mentioned that labour force projections have often been substantially underestimated. Nevertheless, it should be noted that labour force projections are not equivalent to forecast, and have their limitations: (i) If there are changes to the factors influencing population growth and population structure (such as change in the immigration policies), the effectiveness of the projection results will be affected; (ii) If factors having an impact on the labour force in the past are no longer applicable (or the degree of applicability has changed), or some new factors have emerged, the current projection results may not fully reflect the impact of these factors on the LFPR. In this connection, labour force projections should be updated periodically to reflect the impact of the latest policies and socio-economic factors on the labour force.
C&SD will regularly review the past projection results and explore if there is room for further improvement in the projection methodology. Taking the 2001-based and 2003-based labour force projections as examples, the projected labour force (including foreign domestic helpers) as at 2014 was only 2.2% and 2.8% lower than the actual figures, and the average annual growth rates were only underestimated by 0.2% and 0.3% respectively. The projection discrepancy was mainly due to the fact that the launch of the Statutory Minimum Wage (SMW), which was a major policy change, could not be anticipated at the time. The SMW was implemented in May 2011. During the initial period of implementation, the local economic environment was favourable with abundant job opportunities, encouraging many people to join or rejoin the labour market. Thus, there was a substantial increase in the labour force during 2010-2013, at an average annual rate of 2.0%, which was higher than the trend before implementation of the SMW (average annual growth rate was 0.7% during the 10-year period of 2000-2010). As from the last round of labour force projections (2012-based), impact of the SMW has been taken into account.