EDA on Migration Pattern to New Zealand between 2008 to 2022

1. Excecutive Summary

In this report we attempt to answer the question of how immigration pattern has changed before and after the pandemic.

Data between year 2008 to September 2022 are used. There are signs where immigration trend from certain nationality can be correlated to political and economic environment in their country of origin during this period.

India, China and the Philippines are currently the nationalities with the highest number of immigrants.

Great Britain has enjoyed good domestic economic growth for the last decade and there has been a steady decline in migrants from Great Britain.

South African immigrants have dropped until year 2015 when there were a number of social and political movements, then the immigrant number have been rising since.

Post-pandemic, the number of recent resident visa holder returning to New Zealand started increasing rapidly - roughly when New Zealand moved to the COVID-19 Protection Framework, also known as the traffic light system. That trend continues increase sharply and is currently at a historical high point. From the analysis below, a large number of these are likely to be young families returning to New Zealand with their children.

The number of Migrants with work permit and student visa have dropped to 48% and 38% from its peak respectively. There may be early signs of these numbers stabilizing but an increasing trend cannot be seen yet.

Other general observations about gender, nationality and age range are also presented in this report

2. Table Of Content

1. Executive Summary
2. Table Of Content
3. Introduction
4. Methodology
   • 4.1 The Data Set
     • 4.1.1 Gender
     • 4.1.2 Age Range
     • 4.1.3 Nationality
   • 4.2 Exploratory Data Analysis
5. Exploratory Data Analysis
   • 5.1 General Trends In Nationality and Seasonal Migration Trends
   • 5.2 General Trends In Visa Type
   • 5.3 General Trends In Migrant Nationality Versus Visa Type
   • 5.4 General Trends In Migrant Gender
   • 5.5 General Trends In Gender Versus Nationality And Visa Type
   • 5.6 General Trends In Visa Type Versus Gender
   • 5.7 General Trends In Age Range
   • 5.8 General Trends In Age Range Versus Nationality
   • 5.9 General Trends In Age Range Versus Visa Type
   • 5.10 General Trends In Age Range Versus Gender
6. Conclusions

3. Introduction

The analysis in this report examines general pattern of immigration into New Zealand. In particular, we try to examine how immigration trends have changed during and after the COVID-19 pandemic.

Each of the variables gender, visa type, nationality and age range are discussed in relation to one another.

Most migration activities have significantly decreased during the pandemic, until there was a large influx of recent residents returning to New Zealand.

The result of the following analysis may be useful in planning and designing support services for new immigrants or returning residents in settling in an adapting to life in New Zealand.

4. Methodology

4.1 The Data Set

Data source: “Ministry of Business, Innovation and Employment migration dataset”

The data contains 6 columns/variables.

The variables are Date, Visa Type, Gender, Age Range, Nationality and Count.

Variable	Description
Date	the year-month when the number of migrants entering NZ was recorded
Visa Type	visa type the migrants held when entering NZ
Gender	gender of the migrants entering NZ
Age Range	age group of migrants entering NZ
Nationality	original nationality of the migrants entering NZ
Count	number of migrants recorded

The following table provides a summary of the data type and summary statistic of each variable.

It can be seen that the data used in this study was recorded between 2008-07-31 and 2022-09-30.

On any combination of monthly period, visa type, gender, age range, the migrant count could be as little as 0 or as large as 19395.

Table 4.1 Data Summary

	Date	Visa Type	Gender	Age Range	Nationality	Count
	Min. :2008-07-31	Length:688441	Length:688441	Length:688441	Length:688441	Min. : 0.00
	1st Qu.:2012-02-29	Class :character	Class :character	Class :character	Class :character	1st Qu.: 3.00
	Median :2015-09-30	Mode :character	Mode :character	Mode :character	Mode :character	Median : 6.00
	Mean :2015-09-11	NA	NA	NA	NA	Mean : 95.17
	3rd Qu.:2019-03-31	NA	NA	NA	NA	3rd Qu.: 27.00
	Max. :2022-09-30	NA	NA	NA	NA	Max. :19395.00

4.1.1 Gender

There are slightly more male than female migrant in the overall data.

The data points without gender recorded was only 0.48% of the dataset.

Fig 4.1 Examine gender variable

4.1.2 Age Range

The age of the migrants skew toward young age. Majority of the migrants are under 39 years old with most migrants in the 20 to 29 age group for both genders.

Figure 4.2 Examine age distribution and gender

Interestingly, a significant portion of data points with no gender recorded is in the 65+ age group.

Fig 4.3 Examine age distribution where no gender is recorded

4.1.3 Nationality

There are 223 unique nationalities recorded in the data set.

Since 2008, the top migrant nationalities in descending order are India, China, Great Britain, Philippines, South Africa, Fiji, South Korea, Samoa, United States of America and Germany

Fig 4.4 Top migrant nationalities

4.2 Exploratory Data Analysis

The trend of each variable over time (Gender, Age Range and Nationality) are examined.

Each variable and their correlation with each other are also examined.

5 Exploratory Data Analysis

5.1 General Trends In Nationality and Seasonal Migration Trends

Top 10 nationalities are plotted below.

Each year, there is a sharp decline in migrant numbers in December.

Migrant number have declined since the first COVID-19 lockdown at the end of March 2020.

At the beginning of 2020, follownig the expected seasonal trend, migrant numbers from all nationality increased as expected until the March 2020 lockdown.

Migration number from China did not increase up as much as expected at the beginning of 2020 and started seeing a drop in number earlier than other countries. This can be attributed to border closures from China as it ws the first countries to report a COVID-19 outbreak.

Prior to year 2013, Great Britain was the highest source of migrant, followed by China and then India.

Up until the pandemic begins there were steady increase in migrant numbers from India, China, the Philippines and South Africa.

Great Britain on the other hand show steady decrease in migrant numbers since year 2008 at the start of this data set. It has been reported that Great Britain has been enjoying good economic growth in the past decade. This could be related to generally less people emigrating from Great Britain.

Another country showing steady decline in migrant number is Fiji.

South Africa sees decline in migrant number until 2015 and then increased again. There were a number of social and political protest and movements in 2015 in South Africa.

For most of these nationalities, there is a small increasing trend in number of migrants recently.

Fig 5.1 Top 10 nationality of migrants

The number of migrants are the lowest during December each year and are also lower around June, except during the pandemic (2020 - 2021).

Therefore in the following model, a “Year-End” variable is created and is marked as “1” for December records and “0” for others.

“Month” is also extracted from “Date” to account for the seasonal trend.

Fig 5.2 Seasonal trends in immigration

After some trial-and-error, the following polyomial model is found to used to fit the migration trend well. (A step-wise selection of variable could be carried out but this example is relatively simple, trial-and-error gave very good result.)

model <- lm(Count ~ poly(Year,4) + poly(Month, 4) + Year_End, data=data)

Typical dianostic plots to evaluate the quality and validity of a linear regression model are shown below.

1. The model fitted gives randomly scattered residuals suggesting that the model explained most of the trends.

2. The normal quantile-quantile plot shows a straight line except at the lowest and highest ends. The data only deviate from the assumption that the data is approximately normal at the lowest and highest end of the data.

3. The square-root of standardized residual plot shows few data point greater than 1.5, suggesting those are outliers, but there are not many of them.

4. The residual versus leverage plot shows no high leverage outliers that would create strong bias in the model.
   

Fig 5.3 Evaluation of polynomial model for migrant trend

The polynomial model is overlayed onto the actual migration numbers.

The model fit the actual number reasonably well.

Residuals increases in magnitude after early 2020 due to the pandemic.

5.4 Plotting the model prediction against actual data

5.2 General Trends In Visa Type

The number of migrants returnig to NZ, who are recent NZ residents have been going through an approximate 10 year cycle. It was at a peak in 2009 followed by another peak around 2019.

The number of recent residents returning to NZ has been on a downward trend before the pandemic hit.

Toward the end of the pandemic, the number of returning residents greatly increased since November 2021 and is still on an upward trend.

Since the pandemic, the number of migrant moving to NZ on work and student visa have declined sharply. The numbers are still decreasing during 2022.

The number of work visa holders has dropped to approximately 48% from its peak at April 2020.

The number of student visa holders has dropped to approximately 38% from its peak before the pandemic.

There has been a steady increase in the number of people entering with work permit but that number has declined greatly since the panedemic started.

Very little amount of visa types were under the “Other” category.

Fig 5.5 Trend in number of migrants versus visa types

5.3 General Trend in Migrant Nationality Versus Visa Type

5.3.1 Recent Residents

There were significantly less recent residents migrating back to NZ during the pandemic.

The number of migrants have greatly increased in late 2021 for some of these top migrant nationalities. This is around the time when COVID restrictions began to loosen and the “traffic light” system was introduced. For example, Filipino and Indian recent residents moving back to NZ have surpassed pre-pandemic level. Returnign Chinese recent residents have also been increasing.

The number of South African recent residents plateaued during the pandemic but did not decrease. From late 2021, the number of migrants increase sharply, similar to other nationalities.

5.3.2 Students

The number of students have greatly decreased since the pandemic.

There is a slight increase in Chinese students recently.

5.3.3 Work permits

Migrants holding work permits have decreased sharply since the pandemic. Immigration NZ would not have accepted new applications and these migrants would have difficulty entering NZ during the pandemic.

The level of migrants with work permits is still much lower than pre-pandemic level.

5.3.4 Others

Other visa holders are a small portion of this data set.

Fig 5.6 Breakdown of migrant visa types and nationality

5.4 General Trends In Migrant Gender

The number of migrants for both male and female show the same trend over time with male migrants always higher in number.

Fig 5.7 Trend in number of migrants versus gender

5.5 General Trends In Gender Versus Nationality And Visa Type

Indian and the Philippines are the two nationalities that show significantly more male than female migrants.

Fig 5.8 Trend in number of migrants by nationality and gender

Breaking down Indian migrants by visa type and gender shows that most of the difference in migrant number between male and female can be attributed to more male work permit holders than female (Fig 5.9 and Fig 5.10).

Indian male student visa holders are also much higher in number than than female student visa holders before the pandemic.

Fig 5.9 Trend in number of Indian migrants versus Visa Type and Gender

Fig 5.10 Trend in number of Philippino migrants versus Visa Type and Gender

5.6 General Trends In Visa Type Versus Gender

There are more male work permit holders than female work permit holders.

There are slightly more female recent residents than male.

There are slightly more male student visa holders than female.

Fig 5.11 Trend in number of migrants with different visa types and gender.

5.7 General Trends In Age Range

The largest age group of migrants were 20 to 29 years old up until the first pandemic lockdown. Since the pandemic, the 30 to 39 year-old age group has become the age group with the highest number of migrants.

Recently there is an upward trend again in immigration in different age groups, especially 0-19, 20-29 and 30-39.

The 40-49 age group plateaued and only slightly decreased during the pandemic.

In the following plots, we shall examine these trends by breaking down the visa type and nationality more.

Fig 5.12 Trend in number of migrants in differetn age groups

5.8 General Trends In Age Range Versus Nationality

In general, even thoguh there had been a drop in the number of migrants around the age of 30-39 years, in the top 6 countries, the number of immigrants into NZ in this age range remain signficantly high. Except South Africa, in recent months in 2022, the number of immigrants in this age group are the highest after the pandemic, except for migrants from South Africa where the highest number age range is 0-19 years old.

China, Great Britain and India used to have much higher number of migrants in the 20-29 years age range before the pandemic, where NZ would likely to benefit from having these migrants as additions to the NZ workforce. This has dropped signficant during or after the pandemic.

In pre-pandemic era, 20-29 year-old Indian migrants have been the largest Age-Nationality group of immigrants by far. After the pandemic, the number of 30-39 year-old Indian migrants have surpassed the 20-29 year-old age group.

In terms of 0-19 years old migrants, China was on a steady upward trend until the pandemic hits.

During the pandemic even through there were some drop off in Indian, Filipino and South African migrants in the 0-19 years old age range, we are now seeing an increasing trend.

As mentioned above, the 40-49 age range is an interesting group in that in most nationalities, except Great Britain, the number of migrants are relatively stable during the pandemic.

There are few categories where there were a slower increase in the number of immigrants but still continued to grow during and after the pandemic nevertheless. These are Indian 30-39, 40-49 year-olds, Filipino 50-59 year-olds and South African 50-59 year-olds.

Fig 5.13 Trends in number of migrants in different age range and nationalities

It can be seen from the plots below that the increasing migrant numbers for 0-19 years old from India, the Phillipines and South Africa are recent resident holders and not student visa holders.

This is consistent the observation that migrants in the recent resident category has increased dramatically. Also in the next section, we shall also see that the 30-39 years old age range show the sharpest increase after the pandemic. They are likely to the parents of the 0-19 year olds.

Fig 5.14 Trends in number of 0-19 years migrants from India, the Philippines and South Africa

5.9 General Trends In Age Range Versus Visa Type

From the plots below it can be seen that 0-19 years old are mostly like to be student visa holders.

For 20-29 year-olds, there signficant number of student visa holders and also work permit holders.

For 30-39 year-olds, there are significant numbers of student visa holders and work visa holders. The number of work permit holders between 30 and 39 years old have surpassed the 20 to 29 age group since the pandemic started.

30-39 year-olds also have the most number in recent residents category.

As mention above in section 5.5 for the overall trend in visa types. In most age groups (0-59 years), the number of recent resident migranting back to NZ increased since late 2021 but the number of work permit holders started decreasing sharply starting around the same time.

Fig 5.15 Trend in number of migrants in different age range and visa type

5.10 General Trends In Age Range Versus Gender

In general there are more male applicant then females, similar to what is observed in section 5.4.

In the 20-29 and 30-39 age range, there has been small increases recently in female migrant but the same trend is not present in the male data.

Fig 5.16 Trend in number of migrants in different age range and gender

6. Conclusions

• Overall number of migrants moving into New Zealand has decreased and is much lower than pre-pandemic level. It is still at a higher level than around 2010 and there is sign that the number of migrants could be increasing again.

• Majority of post-pandemic migrants are likely to be young families wtih children that are recent residents (30 - 39 year olds with 0 - 19 year old children). The number of these migrants have increased dramatically since Nov 2021.

• The number of work permit holders have dropped significantly. The main source of these migrants are India and the Philippines before or after the pandemic.

• There are generally more male migrants than femail migrants. The visa category that see the most of this difference is in work permit.

• There are early signs of the number of chinese students increasing.