More People Counting Updates
And wishing someone a happy birthday. Numberwang 5
(Artist rendering of the Hong Kong Children's Hospital, Source: Cundall)
Hong Kong's population might have gotten the spotlight in new data released by the Census and Statistics Department in the last week, but there have been other demographic data updates as well, so let's get straight to it.
Prayers answered! Births, Deaths, and Marriages
Back in Numberwang 1 I lamented the end of Hong Kong's Monthly Digest of Statistics, as it was the only online source that contained crucial demographic data: births, deaths, and marriages. So imagine my surprise when the Hong Kong Free Press -- please support and donate -- published an article that there were fewer than 30,000 births in the 12 months up to June 2026.
According to figures released by the Census and Statistics Department on Tuesday, a total of 29,709 babies were born from July 2025 to the end of June this year.
How in the world did they get the exact figure? It turns out the Census and Statistics website has added three additional data tables online!
The tables were added on August 18, 2026, the same day that the preliminary population figures for mid-2026 were released.
A quick scan and comparison of the new births data reveals that the numbers in these new tables are completely different than the data provided by the Monthly Digest of Statistics. The reason? The Monthly Digest data was obtained by the Immigration Department, which is responsible for the registration of births, so they provided the number of registrations each month, not the actual number of babies born that month. (For example, when the Covid-19 pandemic was at its worst, many new parents decided to stay home and wait weeks before they would register their baby for fear of catching Covid.) The new data? It provides the actual number of people born that month; which is an improvement and it's awesome.
With new data in hand, we can update one of the charts posted in Numberwang 1: the number of monthly births before and after the HK$20,000 baby bonus scheme was implemented. The old and updated graphs make an interesting comparison.
The new graph confirms that the visible increase in births due to the baby bonus scheme lasted just six months! It probably wouldn't have mattered, but how dumb is it to announce a baby bonus scheme and then completely ignore it with no further promotion after just a few weeks? It certainly felt as if they were more concerned with cost savings rather than the aging population.
Incidentally, the 2,276 births in June 2026 is believed to be the second-lowest on record since World War 2; in November 2022, the number of births was 2,275.
With 14,469 births in the first six months of 2026, does this mean Hong Kong will have fewer than 30,000 births for all of 2026, for the first time since World War 2? Not necessarily. Since 1976, Hong Kong has had more births in the second half of the year compared to the first half (a 46.5-53.5 split), and only twice in those 50 years has there been more January-June births than July-December births: 2020 due to Covid and 2025 due to the baby bonus scheme. Watch this space in 6 months.
To wrap up this section, here are some shiny new graphs on deaths and marriages. Feel free to draw your own conclusions.
Death Expectancy and Death Tables. Viewer discretion advised.
Life expectancy at birth in 2025 rose slightly compared with 2024, and remained among the highest in the world, according to preliminary data from the Census and Statistics Department.
Male life expectancy went up 0.6 years, from 82.7 last year to 83.3 this year.
Female life expectancy is at an all-time high, at 88.7 years, compared to 88.2 last year.
While we're at the topic of life expectancies, here is an important publication by the Census and Statistics Department that has a cheery-sounding name, but it's quite morbid in reality: The Hong Kong Life Tables, 2025. Life expectancy at birth is determined using the Hong Kong Life Tables, and I've written a postscript below for those who wish to learn how it is computed in more detail.
Life tables for countries are crucial because they give an idea of how long people tend to live and their chances of death as they grow older, so in a sense they really should have named this chart Hong Kong Death Tables. Anyway, you can probably deduce that insurance companies use Life Tables to determine premiums and annuity payments, but Life Tables are also helpful for governments, pension funds, public health researchers, demographers, statisticians, and economists. Not only does the CSD publish the Life Tables for 2025, they also predict future Life Tables!
Life tables are of use when planning across the entire population, and less useful when predicting how long one particular person will live. The Google spreadsheet below has the latest Life Table for Hong Kong based on 2025 (preliminary) data.
To read this chart, let's use a particular age as an example. It turns out Regina Ip turns 76 on August 24, the day this blog entry was posted, so let's look up the numbers for women aged 76:
From here, we can make the following observations from the chart based on the data:
Consider a hypothetical group of 1,000,000 girls born in Hong Kong who, throughout their lives, experience the death rates recorded for Hong Kong women in 2025. According to the table, about 881,934 of them would reach their 76th birthday (i.e. 118,066 would have already passed away),
A woman who has just turned 76 has on average a 1.14% probability of dying before her next birthday, i.e. on average a 98.86% probability of being able to celebrate her 77th birthday, therefore,
Around 10,058 out of those 881,934 remaining women would die in the next year before their 77th birthday, and finally,
The expected remaining lifespan of women at age 76 is 16.29 years.
It's important to note that these numbers do not apply to Regina herself, but rather they describe average outcomes for a hypothetical group of women who have just reached age 76 and experience Hong Kong's 2025 mortality rates. Anyway, happy birthday, Regina! Live life to the fullest!
How is Hong Kong's life expectancy at birth computed?
The most important pieces of information required are:
• How many people of each age are living in Hong Kong, and
• How many people of each age are dying in each year in Hong Kong
(both of which are tracked by the Census and Statistics Department).
Using that information, the CSD estimates the probability that a person of each age will die before their next birthday. However, for very young ages and very old ages, the probabilities are relatively more difficult to figure out for various reasons (for example, the exact ages of the elderly may not be as reliable), so what statisticians do is use mathematical modeling to come up with their best estimate. Based on various assumptions of the population's risk of death for young children and for people who are very old, the CSD uses what it thinks to be the best model by age. Explaining those adjustments is beyond the scope of this blog.
So the algorithm generally works like this:
1. We imagine a group of one million newborn babies in the same year.
2. We estimate how many of them will die before their next birthday.
3. We carry over the survivors forward to the next age.
4. We go back to step 2 and repeat the process, until we have an estimate of how many people aged 84 are still alive.
5. At age 85 and older, we use modeling to extrapolate rates rather than using observed rates, all the way until there are no more survivors. (As you might imagine, there are different modeling methods to make such an estimate, and this can lead to some dispute as to whether a particular method used by the CSD is actually best.)
6. Expected life expectancy at birth is simply the number of years the one million people lived, then divided by one million. (For nearly all ages, we assume deaths are spread evenly throughout the year, so someone who dies at age d has lived d + 0.5 years. Age 0 receives special treatment, as a vast majority of deaths occur within the first 28 days. For ages 100 and higher, the published table combines everyone aged 100 or above into one row, the CSD continues the modeled calculations age by age until no survivors remain, then adds together all the years lived after age 100. If you've looked carefully at the 2025 Life Tables, you'll notice the chart only goes up to 100; in the last row, the chart indicates the average life expectancy of someone who turns 100.)
As mortality rates differ by sex, we run the algorithm separately for boys and girls. So based on the 2025 Hong Kong Life Tables above, the imaginary group of one million girls lives a combined 88.73 million years, so life expectancy at birth is 88.73 million ÷ 1 million people = 88.73 years.
(My thanks to Prof. Ben Cowling for his assistance with today's blog.)
