|
1. |
Introduction |
|
1.1 |
A press conference was held on 11 August 2000 to announce the revision to the method of compiling population estimates of Hong Kong. This article explains the rationale for the revision and introduces the concept of the "Hong Kong Resident Population". |
|
2.
|
Previous method of compiling population estimates
|
|
2.1 |
All along, the so-called "extended de facto" method has been used by the Census and Statistics Department (C&SD) to compile population estimates. Under a "de facto" concept, the population includes all persons who are in Hong Kong at the reference time-point, i.e. this method is equivalent to taking a 'snapshot' of the population at a reference time-point. The population includes Hong Kong Permanent Residents, Hong Kong Non-permanent Residents and visitors. |
|
2.2 |
"Extended" relates to the fact that for a Hong Kong Permanent Resident, he/she will still be counted as part of the Hong Kong Population if, at the reference time-point, he/she is not in Hong Kong but temporarily in the Mainland or Macau. |
|
2.3 |
The application of the "extended" compilation method is intended to avoid fluctuations in the "de facto" population figure around major public holidays when the movement of people between Hong Kong and Mainland/Macau is enormous. And in the past, most such departures to Mainland/Macau were of a temporary nature. |
|
2.4 |
In practical terms, population estimates are compiled as described below. |
Population estimate at a "census moment" based on the "extended de facto" method |
|
2.5 |
A population base is established at each population census (or by-census) moment. In addition to those who are physically in Hong Kong, also included are those Hong Kong residents who are working in Mainland/Macau and those who are usually living in Hong Kong but in Mainland/Macau for short trips at the census moment, and are reported by family members in Hong Kong as such. |
Population estimate at any reference time-point in between census moments based on the "extended de facto" method |
|
2.6 |
The population estimate at any reference time-point (t) is obtained by the following formula: |
|
|
| Population (at time-point t) |
|
= |
Population (at census moment c, based on the concept of the "extended de facto" method) |
|
|
+ births (from c to t) |
|
|
- deaths (from c to t) |
|
|
+ arrivals (from c to t) |
|
|
- departures (from c to t) | |
|
|
and "arrivals - departures" excludes departures and arrivals of Hong Kong Permanent Residents to/from Mainland/Macau during the period from c to t. The implementation of this exclusion is to deal with the situation that a great deal of departures to Mainland/Macau shortly before the reference time-point may be related to short-term stays (though these would be matched with arrivals back in Hong Kong shortly after the time-point) and may consequently lead to highly fluctuating population estimates if they are all included. |
|
3.
|
Limitations of the "extended de facto" method
|
|
3.1 |
Since the residency and mobility patterns of the Hong Kong Population are undergoing change, the "extended de facto" method has gradually become less appropriate. The main reasons are : |
|
|
(a) |
The trend for Hong Kong Permanent Residents to stay in the Mainland or Macau on a long-term basis for reasons of retirement, work or marriage is on the increase. Also, a considerable number of Hong Kong residents are staying overseas for work, touring or studies at the reference time-point. Applying the "de facto" method would lead to fluctuations in the population estimates and limit its ability to reflect the actual situation. |
|
|
(b) |
According to the definition of "de facto" population, visitors to Hong Kong are included in the Hong Kong Population. But strictly speaking, visitors are not part of the Hong Kong Population. |
|
|
(c) |
The existing compilation method makes use of the difference between the number of arrivals and the number of departures. As those numbers are large, their difference may not be stable at times, especially around the commonly used reference time-points, e.g. end-June and end-December. |
|
3.2 |
In regard to (a) above, the application of the extended de facto method leads to some over-estimation of the population because indeed there are people who go to Mainland/Macau for long-term stay or even for good (rather than just for short trips), and the trend is for this to intensify. It should be noted, though, that the population estimate released about 2 months after a reference time-point is, strictly speaking, provisional in nature. |
|
3.3 |
This is because, when an ensuing population census (or by-census) is conducted and the new population base is established, those who have indeed gone to Mainland/Macau for long-term stay or for good would not be counted and for the population estimates in respect of the intervening reference time-points, a "retrojection" process will be carried out, with the number of those people deducted. In other words, the provisional population estimates are finalized after the next population census/by-census has been taken. (It has to be noted, though, that the population census/by-census performs the benchmarking function, such that the finalization process which involves revision of estimates is not solely due to the above factor.) |
|
3.4 |
However, the finalization takes place years after the individual reference time-points. For example, the estimate for mid-1992 was "finalized" only in the latter part of 1996 when data from the 1996 By-census (March 1996) became available, and hence a time lag of slightly more than 4 years. (And for the estimate for mid-1993: a time lag of slightly more than 3 years; for mid-1994: 2 years; and so on.) |
|
3.5 |
Previously, when the volume of traffic and the number of people going to Mainland/Macau for long-term stay (or for good) were not so large, the problem was limited as the extent of "revision" would also be relatively small. That is, the "provisional" estimates could be used in the interim (a rather long interim period though) without too much concern. The situation is, however, changing. |
|
4.
|
Merits of changing over to the "resident population" method for compiling population estimates
|
|
4.1 |
In the early days, many countries and territories applied the "de facto" method for compiling population estimates. The main reason is that the compilation procedure is simple and the required data are easier to obtain. With statistical development and advancement, advanced countries and territories have gradually changed over to using the "resident population" method. |
|
4.2 |
In changing over to using the "resident population" method for compiling population estimates of Hong Kong, the following considerations are relevant: |
|
|
(a) |
The residency and mobility patterns of the Hong Kong Population
Recent studies have shown that there have been notable changes in the residency and mobility patterns of the Hong Kong Population. The application of the "resident population" method could better reflect the actual situation. |
|
|
(b) |
International trend
Many countries and territories have changed over from the "de facto" method to the "resident population" method for compiling population estimates. Hong Kong should also follow this international trend. |
|
|
(c) |
Statistical conceptualization
Using the "resident population" concept to measure the population size of a place is actually more relevant from the statistical theory standpoint. Its previous non-application was mainly due to limitations of insufficient data. Also, in compiling various socio-economic statistical indicators (e.g. birth rate and death rate), the use of the "resident population" concept is generally more meaningful. |
|
|
(d) |
Development of information technology
The computational capabilities of computers now available have enhanced significantly. Large volume of anonymized passenger movement data can be handled, thus enabling the computation of "resident population" figures at the reference time-point. |
|
5.
|
Revising the method for compiling population estimates
|
|
5.1 |
Concluding from the results of relevant studies, it is considered that Hong Kong has met the necessary requirements of applying a compilation method which is based on the concept of "resident population". In doing so, Hong Kong follows the statistical practice of other advanced economies. |
|
5.2 |
"Resident population" is a clear-cut concept in international statistical standard but the practical definitions adopted vary from place to place, as the residency and mobility patterns unique to each place need to be given adequate consideration. International statistical organizations have pointed out in particular that, owing to business and social development, the "mobility" of residents of certain countries/territories is rather high. In handling the population statistics of these countries/territories, the appropriate authorities should consider the situation in depth. |
|
5.3 |
In the case of Hong Kong, our studies show that the "resident population" of Hong Kong (which we refer to as the "Hong Kong Resident Population") should be defined to include "Usual Residents" and "Mobile Residents". |
|
5.4 |
In consideration of issues relating to concepts and availability of data, the technical definition of the "Usual Residents" would include two categories of people: (1) Hong Kong Permanent Residents who have stayed in Hong Kong for at least three months during the six months before or for at least three months during the six months after the reference time-point, regardless of whether they are in Hong Kong or not at the reference time-point; and (2) Hong Kong Non-permanent Residents who are in Hong Kong at the reference time-point. |
|
5.5 |
As for "Mobile Residents", they are Hong Kong Permanent Residents who have stayed in Hong Kong for at least one month but less than three months during the six months before or for at least one month but less than three months during the six months after the reference time-point, regardless of whether they are in Hong Kong or not at the reference time-point. |
|
5.6 |
The amount of time of stay in Hong Kong of "Mobile Residents" is less than that of the "Usual Residents". Nevertheless, the "Mobile Residents" have a close link with Hong Kong and most probably they have a regular residence in Hong Kong and utilize much of Hong Kong's facilities and services. In this regard, they should be considered as part of the Hong Kong Population. |
|
5.7 |
Under the new method for compiling population estimates, a population census/by-census continues to serve the benchmarking function. Updating of the population estimates to any new reference time-point, however, does not depend on just immigration control-point data on the balance of total arrivals over departures in regard to population movement. Instead, in respect of Hong Kong Permanent Residents, arrival and departure records of individual persons are linked for statistical computation (albeit the records are anonymised and un-identifiable to the individuals). For Hong Kong Non-permanent Residents, the balance of arrivals over departures is still used, since the record-linkage method on individuals cannot be applied to them given the form of their arrival/departure records. Also, they are grouped under "Usual Residents". This is because for the duration that they hold that status of "Non-permanent Resident" they can be expected to be usually staying in Hong Kong. |
|
5.8 |
Below is an illustration of the situations of different Mobile Residents: |
|
|
(1) |
Persons staying 5 to 6 days each week in the Mainland and staying regularly in Hong Kong during the weekends. As at end-1999, there were some 56 000 such persons. They are mainly those working in the Mainland and return to Hong Kong to stay with their families during the weekends. |
|
|
(2) |
Persons staying for a major proportion of time in Mainland/Macau and returning to Hong Kong from time to time. There were some 72 000 such persons at end-1999. Presumably they maintain residences in both Mainland/Macau and Hong Kong, and travel frequently between these places for business, work or family reasons. |
|
|
(3) |
Elderly persons (aged 60 or over) staying for a major proportion of time in Mainland/Macau. There were some 7 000 such persons. Many of them have retired and settled in the Mainland but come back to Hong Kong frequently to visit children or relatives. |
|
|
(4) |
Persons in school-attending age returning from overseas to stay in Hong Kong for several periods of time in a year. There were some 21 000 such persons. These are mainly Hong Kong students studying abroad and spending their vacations in Hong Kong. |
|
|
(5) |
Besides, there are persons staying in overseas countries/territories but also staying frequently in Hong Kong for business, work or family reasons. There were some 32 000 such persons at end-1999. |
|
5.9 |
Chart 1 provides a diagrammatic presentation of the residency/mobility profiles. |
|
5.10 |
As at end-1999, there were a total of 188 000 "Mobile Residents". The above figures have been obtained through statistical processing of anonymized passenger movement data. The compilation procedures are conducted through computer processing, with all the personal data being anonymous and the identities kept strictly confidential. |
|
6.
|
Comparison of the old and the new population estimates
|
|
6.1 |
On the basis of the new method for compiling population estimates, the "Hong Kong Resident Population" as at end-1999 is 6.76 million, which is less than the previously released "extended de facto" population by 0.21 million (i.e. 3%). However, one should note that the previously released figure (6.97 million) includes 0.14 million visitors, who are not included in the figure of "Hong Kong Resident Population". |
|
6.2 |
Table 1 presents the population estimates from 1996 to 1999, compiled based on the old and the new methods. Since the required data are not available for earlier dates, the estimates of "Hong Kong Resident Population" in the statistical series are only provided in respect of reference time-points from 1996 onwards. |
|
6.3 |
A comparison of major statistical indicators between the extended de facto population and the Hong Kong Resident Population is shown in Table 2. |
|
7.
|
Implications of the application of the revised compilation method for the Government's planning work
|
|
7.1 |
Replacing application of the "de facto" method by the "resident population" method for compiling population estimates is a technical revision. It may be noticed that the changes in the figures are not very large. |
|
7.2 |
In the short run, the revised series of population estimates will not have significant effect on the planning of public services. |
|
7.3 |
In the long run, as the revised series will better reflect the actual situation of the Hong Kong Population and provide richer information on the residency and mobility patterns, they will contribute better to government's long term planning work. |
|
8.
|
Implications of the application of the revised compilation method for labour statistics
|
|
8.1 |
As regards labour statistics, the unemployment and underemployment rates are not affected by the use of the revised method of compiling population estimates. |
|
8.2 |
As for the size of the labour force and the number of employed, unemployed and underemployed persons, there are downward adjustments of up to about 5% (e.g. taking the number of unemployed persons as 180 000, the downward adjustment is about 9 000). |
|
8.3 |
The extent of revision is different from that for the overall population since there is also some adjustment to the sex-age composition of the population (see Table 3). But it should be noted that the unemployment rate, being a ratio between number of unemployed persons (U) and the labour force (L), is generally not changed since U and L are both adjusted to a very similar extent. |
|
9.
|
Release of population estimates
|
|
9.1 |
The application of the "resident population" method requires information on the amount of time of stay in Hong Kong of residents during the six months before and after the reference time-point. It follows that the population estimates could not be obtained within one or two months after the reference time-point. |
|
9.2 |
In order to provide relevant figures as early as possible for use, the C&SD will release "provisional estimates" for reference within two months after the reference time-point. Revised estimates will be released a further six months later. |
|
9.3 |
That is to say, population estimates will be released every half-year, for the mid-year and year-end positions, by way of press release as in the past. To provide a concrete example, the provisional estimates for mid-2000 were released on 22 August 2000, while the revised estimates will be released in mid-February 2001. |
|
9.4 |
The latest population estimates are shown in Table 4. |
|
9.5 |
Chart 2 and Table 5 present data series in respect of population estimates and population growth rates for 1990-2000 for ease of reference. |
|
10.
|
Population Projections
|
|
10.1 |
With the new method of compiling population estimates in place, our next step will be to compile population projections for the future. These estimates will be released in a few months. |
|
11.
|
Concluding remarks
|
|
11.1 |
The degree of difficulty in compiling population estimates differs among countries/territories. In comparison with other places, statistics on birth and death are easily obtained in Hong Kong. However, it is not so easy with regard to residency and mobility. Since a resident is not required to state the time and purpose of staying in or being away from Hong Kong at arrival or departure, and the traffic flow is massive (over a hundred million movements a year), data handling is of a rather complex nature and detailed study and analysis is required to enable reliable figures to be compiled. |
