(Sorry these charts below are confusing. Please see the new charts of my second attempt.)

San Francisco is losing children population. Family move out of the city because of bad schools and expensive housing. Such is the popular conception of the issue of living in the city. But I have also heard some contradicting figures, like the kindergarten enrollment is steadily increasing in recent years. So rather than listening only to anecdotes, I am most interested in looking for some hard data to understand the extent of the issue.

Luckily we have the wonderful U.S. Census Bureau. They have collected and published lots of data at great level of details. Not only is the census of 1990, 2000 and 2010 available, the American Community Survey has also provided 1-Year estimates for every year since 2000. This allows me to do an analysis on the demographics. By tracking individual cohort and comparing their population change over time, it is able isolate other factors to provide insight on whether or how children leave the city.

The analysis mostly validate the observation that children are leaving the city before entering school. Moreover it suggests that the initial elementary school enrollment is the only event that triggers mass migration. Other factors, like real estate concerns, may not play any significant role.

First of all the overall population of San Francisco is growing slowly. It has recovered from the lost after the dot-com crash and arrive a near all time high 0f 800,000. This is 11% over the population of 1990.

The children population, in this study it is defined as 0-19 year old, has shown similar fluctuation around the dot-com crash. However the overall trend is slightly down, it is 2% less than 2000 and 6% less than 1990. Since the overall population is up, as the percentage of population the decline is even more steep.

Looking at the population of the group from new born to 4, it has instilled some hope. The number is holding up well and perhaps even on a upward tread. The number is slightly less than 1990 but 11% more than 2000.

The next 2 charts are perhaps the most interesting in this analysis. Rather than looking at one aggregated number, we track the individual cohorts and compare its change in 5 years. This directly answers the question whether children are leaving San Francisco. For example, the blue data point at the lower right hand corner represent the group of 0-4 years in 2005. It says that in 2010, when this cohort turns 5-9, their population is only 72% compare to where they have started. One quarter of them has left due to net migration. The green data point at the upper right hand corner tells a more positive story. It says that the 10-14 cohort in 2005 gains 20% in 5 years when they have turned 15-19 in 2010. Older children are actually moving into the city.

The population in separated in three groups (as compiled by the census bureau), under 5, 5-9 and 10-14. They shows a rather divergent outcome. Nearly all the losses come from the under 5 group. And it is at a rather significant drop of one quarter in size! The lost ease up for the older group. In fact more recent census shows that the cohort actually grow in size!

(One note about the five year change charts, the child mortality rate is negligible here. The changes mainly come from net migration.)

While it is a good for population to grow due to immigration, we are particularly interested to see how well we are retaining children who are already in the city. For this we compare the population of each cohort as a percentage of population instead. The overall pattern remains similar, although it makes the lost of the under 5 group even more terrible. This metrics maybe a little bias against the children population though, because I believe the recent influx of immigrant is more likely to have a higher percentage of childless adult as they are more mobile.

The sharp drop of children after they turn 5 points to one single issue, family are leaving at the time when the child first enroll in school. People are concern about the quality of public schools. The assignment system leave them feel out of control. Many bait out when they are not assigned to popular schools.

A sharp drop of children at age of 5 is a more acute problem than if the drop is gradual over time. It lowers the basis for the older groups. Even when there are population increase in older group, it is already on a smaller base. And it comes too late to make a long term boost to the children population.

On the other hand, this analysis largely discount other issues such as housing problem. As children grow, they need more space as they grow physically bigger, acquired more stuff and become more independent. The housing issue should have become more pressing over time. However the population drop off ceased and it even turned into growth for older groups. Housing issues seems far less influential compare to school enrollment.

To halt the exodus of family, the key is in improving the initial school enrollment. As a parent who are right at this watershed stage, I find the public school quality issue is often a perceived one. Nearly all the schools I have toured leave me a positive impression. And those are not even trophy schools people are clamoring, but simply neighborhood schools. It is unfortunately to find so many people dismiss the schools as unacceptable before even trying.

Here is the data I have prepared for anyone who want to validate or further analyze the result.

2011.06.23 [San Francisco] - comments