N ational T ransfer A ccounts Data Review (Hands On) Amonthep Chawla East-West Center & Nihon University Population Research Institute.
Post on 16-Dec-2015
220 Views
Preview:
Transcript
NNational ational TTransfer ransfer AAccountsccounts
Data Review (Hands Data Review (Hands On)On)
Amonthep ChawlaAmonthep Chawla
East-West CenterEast-West Center& Nihon University Population & Nihon University Population
Research InstituteResearch Institute
National Transfer Accounts
ObjectivesObjectives
► To understand your surveys: your surveys To understand your surveys: your surveys may not provide all information needed to may not provide all information needed to construct NTAconstruct NTA Understand purposes of surveysUnderstand purposes of surveys Review questionnairesReview questionnaires Any missing information on income , expenditure Any missing information on income , expenditure
and demographic characteristicsand demographic characteristics► To aware that you may encounter To aware that you may encounter
estimation problems due to outliers, missing estimation problems due to outliers, missing data, sampling weights, smoothing, etc.data, sampling weights, smoothing, etc.
► To prepare survey data before estimating To prepare survey data before estimating age profiles necessary for NTAage profiles necessary for NTA
National Transfer Accounts
OutlineOutline
► I. Demographic variablesI. Demographic variables► II. Data cleaning II. Data cleaning ► III. Components and classifications of III. Components and classifications of
expendituresexpenditures► IV. Components of income IV. Components of income ►V. SmoothingV. Smoothing
National Transfer Accounts
I. Demographic VariablesI. Demographic Variables
► Age of individuals: what is the age limit?Age of individuals: what is the age limit?► Dummy for household headshipDummy for household headship► Dummy for school attendance (currently Dummy for school attendance (currently
attending school or not)attending school or not)► Education level for school attendance (if Education level for school attendance (if
currently attending school, what is the level currently attending school, what is the level of education: primary, secondary or higher of education: primary, secondary or higher education)education)
► Health characteristics (number of times Health characteristics (number of times visiting hospitals, health usage or any visiting hospitals, health usage or any information regarding to health condition)information regarding to health condition)
► Sampling weights (always use weight option)Sampling weights (always use weight option)
National Transfer Accounts
Household HeadsHousehold Heads
► How household heads are defined: economic How household heads are defined: economic head or self-reported head?head or self-reported head?
► Each household should have one person as a Each household should have one person as a household headhousehold head Two or more household heads: pick the oldest or Two or more household heads: pick the oldest or
the highest income, depending on culture or normthe highest income, depending on culture or norm No head: assign one personNo head: assign one person
► What is the minimum age for head?; should What is the minimum age for head?; should he/she be older than 15? Need to understand he/she be older than 15? Need to understand laws and culture of the countrylaws and culture of the country
National Transfer Accounts
Education: Enrolment RatesEducation: Enrolment Rates
► Plot enrolment rates by age and by education Plot enrolment rates by age and by education level level
► How does the result differ from other surveys, How does the result differ from other surveys, i.e. census and other government document?i.e. census and other government document?
► Any information on private or public school Any information on private or public school enrolmentenrolment
► Check minimum and maximum age of school Check minimum and maximum age of school enrolmentenrolment
► Is there minimum age of school enrolment for Is there minimum age of school enrolment for the kindergarten level? the kindergarten level?
► Be aware that childcare is not educationBe aware that childcare is not education
National Transfer Accounts
II. Data CleaningII. Data Cleaning
►Need to carefully check survey data in Need to carefully check survey data in order to avoid or detect possible data order to avoid or detect possible data problem by problem by summarizing data, using sampling weights, to summarizing data, using sampling weights, to
check for missing data, outliers, data check for missing data, outliers, data underestimating or overestimating, etc.underestimating or overestimating, etc.
Using scatter plots to detect outliers: be aware Using scatter plots to detect outliers: be aware that some extremely high health consumption that some extremely high health consumption (usually by the elderly) may be real(usually by the elderly) may be real
►Missing data: replaced with 0 or imputed, Missing data: replaced with 0 or imputed, depending on how serious this problem isdepending on how serious this problem is
National Transfer Accounts
III. Components and III. Components and Classifications of ExpendituresClassifications of Expenditures► Two types of expenditure: consumption and Two types of expenditure: consumption and
non-consumption expenditurenon-consumption expenditure► Avoid using non-consumption expenditure to Avoid using non-consumption expenditure to
estimate NTA private consumptionestimate NTA private consumption► Examples of non-consumption expenditures are Examples of non-consumption expenditures are
transfers (i.e., taxes, gifts or other transfers made to transfers (i.e., taxes, gifts or other transfers made to other households)other households)
asset income (negative asset income such as interest asset income (negative asset income such as interest expense on mortgage or consumer debt) expense on mortgage or consumer debt)
► Be cautious if using individual consumption Be cautious if using individual consumption rather than household consumption, i.e. missing rather than household consumption, i.e. missing datadata
National Transfer Accounts
Per Capita Private Consumption, J apan, 2004
0
50,000
100,000
150,000
200,000
250,000
0 10 20 30 40 50 60 70 80 90+
Age
Yen
Total Private Consumption
Other
Health
Durable
HousingEducation
Including non-consumptionOnly consumption
Double-hump consumption is more outstanding when non-consumption is included
National Transfer Accounts
Income TaxIncome Tax
► How to estimate income tax at the individual How to estimate income tax at the individual level: 2 methodslevel: 2 methods 1. Using household tax expenditures and 1. Using household tax expenditures and
allocating to each members proportionally to their allocating to each members proportionally to their taxable incometaxable income
2. Not using household tax data but using 2. Not using household tax data but using individual taxable income directly to measure individual taxable income directly to measure individual tax paymentsindividual tax payments
► What are the tax rates in these two methods? What are the tax rates in these two methods? ► The first method allows for household level The first method allows for household level
variation in the tax rate while the second variation in the tax rate while the second method assumes that all household income is method assumes that all household income is taxed at the same ratetaxed at the same rate
National Transfer Accounts
Income Tax (cont.)Income Tax (cont.)
►Method 1 is preferred as it gives better Method 1 is preferred as it gives better information on who actually pay taxesinformation on who actually pay taxes
►Be cautious: income tax data may be Be cautious: income tax data may be underestimated or not reportedunderestimated or not reported
► If possible, check both methods and If possible, check both methods and examine how different the age profiles examine how different the age profiles of income tax based on two methods of income tax based on two methods areare
National Transfer Accounts
Per Capita Income Tax Age Profile, Japan, 2004
0
50,000
100,000
150,000
200,000
250,000
300,000
350,000
400,000
450,000
0 10 20 30 40 50 60 70 80 90+
yen
Using household income tax and allocating to each member based on their taxable income
(earnings, self-employed income and property income)
Not using household tax data
Check: On what types of income is income tax levied in your economy?
Not using household tax data leads to overestimate taxes paid at older ages in Japan
National Transfer Accounts
IV. Components of IncomeIV. Components of Income
►Household surveys report income either at Household surveys report income either at the individual level or the household levelthe individual level or the household level
►Unit of analysis for NTA is at the individual Unit of analysis for NTA is at the individual level: individual income is neededlevel: individual income is needed
►Assignment of incomeAssignment of income labor income: assigned to individual memberslabor income: assigned to individual members asset income and transfer income: assigned to asset income and transfer income: assigned to
household headhousehold head
National Transfer Accounts
IncomeIncome AssignmentAssignmentWage or EarningsWage or Earnings IndividualIndividual
Labor’s share of self-Labor’s share of self-employed (or family employed (or family business) incomebusiness) income
Individual (household Individual (household income is distributed to income is distributed to individuals using NTA individuals using NTA method)method)
Taxes, pension, and other Taxes, pension, and other cash transfers from the cash transfers from the governmentgovernment
IndividualIndividual
Capital’s share of self-Capital’s share of self-employed incomeemployed income
Household headHousehold head
Cash transfers from other Cash transfers from other householdshouseholds
Household headHousehold head
Property incomeProperty income Household headHousehold head
National Transfer Accounts
Property IncomeProperty Income
►Possible components of property Possible components of property income in surveys: interest, rent and income in surveys: interest, rent and dividends; some surveys include dividends; some surveys include property income with non-labor incomeproperty income with non-labor income
►Tabulate the age profile of property Tabulate the age profile of property income (assigned to household head) income (assigned to household head) for estimatingfor estimating corporate income tax, andcorporate income tax, and operating surplus of corporations operating surplus of corporations
National Transfer Accounts
V. SmoothingV. Smoothing
►Visually check if the beginning and Visually check if the beginning and ending of the smoothed and ending of the smoothed and unsmoothed age profiles are differentunsmoothed age profiles are different
►Not over-smoothing: measure Not over-smoothing: measure percentage errors between smoothed percentage errors between smoothed and unsmoothed profiles; choose span and unsmoothed profiles; choose span or bandwidth that yields the least or bandwidth that yields the least percentage error between the twopercentage error between the two
National Transfer Accounts
Some Rules for SmoothingSome Rules for Smoothing
► Smooth and aggregate: smooth components Smooth and aggregate: smooth components of the age profiles and aggregate them up, of the age profiles and aggregate them up, i.e. smooth earnings and labor’s share of i.e. smooth earnings and labor’s share of self-employed income and add them up to self-employed income and add them up to measure the age profile of labor incomemeasure the age profile of labor income
► Do not smooth public and private education Do not smooth public and private education age profilesage profiles
► Use unsmoothed health consumption for Use unsmoothed health consumption for newborns because it is usual to have high newborns because it is usual to have high health consumption at this agehealth consumption at this age
National Transfer Accounts
Rules for Smoothing (cont.)Rules for Smoothing (cont.)
► Income, income taxes and familial Income, income taxes and familial transfer outflows: begin smoothing from transfer outflows: begin smoothing from the adults, excluding those younger age the adults, excluding those younger age group who do not earn income. group who do not earn income. However, problem arises when some However, problem arises when some
beginning age group may appear to have beginning age group may appear to have negative values for these variables. negative values for these variables.
This could be solved by replacing the This could be solved by replacing the negative by the unsmoothed values for the negative by the unsmoothed values for the beginning age group. beginning age group.
National Transfer Accounts
Per Capita Private Education Consumption, Japan, 2004
0
100,000
200,000
300,000
400,000
500,000
600,000
0 10 20 30 40 50 60 70 80 90+
ye
n unsmoothed
smoothed
Smoothed education consumption could eliminate age specific education consumption, i.e., primary school and college education expenses
National Transfer Accounts
Per Capita Private Health Consumption, Japan, 2004
0
20,000
40,000
60,000
80,000
100,000
120,000
140,000
160,000
180,000
200,000
0 10 20 30 40 50 60 70 80 90+
ye
n
Unsmoothed
span=0.01
span=0.1
Different spans yield different results: always check for the variations among young and old ages
Use unsmoothed health consumption for newborns
National Transfer Accounts
Percentage Difference between Smoothed and Unsmoothed Private Health Consumption, Japan, 2004
-14
-12
-10
-8
-6
-4
-2
0
2
4
6
20 25 30 35 40 45 50 55 60 65 70 75 80 85
Per
cen
tag
e D
iffe
ren
ce
span=0.1
span=0.01
Lower spans cause lower percentage difference between smoothed and unsmoothed age profiles than higher spans
National Transfer Accounts
ConclusionsConclusions
► It is important to examine demographic It is important to examine demographic characteristics, particularly education and health, characteristics, particularly education and health, reported in the surveysreported in the surveys
► Need to cleanup data set to avoid or detect data Need to cleanup data set to avoid or detect data problem, which could cause serious mistakes while problem, which could cause serious mistakes while estimating age profiles for NTAestimating age profiles for NTA
► Avoid including non-consumption expenditure while Avoid including non-consumption expenditure while estimating consumption age profile estimating consumption age profile
► Labor income is assigned to individuals while asset Labor income is assigned to individuals while asset income and transfer income is assigned to headincome and transfer income is assigned to head
► Choose appropriate span or bandwidth that causes Choose appropriate span or bandwidth that causes the least variation between smoothed and the least variation between smoothed and unsmoothed age profilesunsmoothed age profiles
top related