Projecting NTAs Tim Miller Center for the Economics and Demography of Aging, UC Berkeley tmiller@demog.berkeley.edu Third NTA Workshop Honolulu, Hawaii.
Post on 01-Apr-2015
215 Views
Preview:
Transcript
Projecting NTAs
Tim MillerCenter for the Economics and
Demography of Aging, UC Berkeley tmiller@demog.berkeley.edu
Third NTA WorkshopHonolulu, HawaiiJanuary 20, 2006
Thanks to the United Nations CELADE for support in developing these models.
Plans for next year (next 3 years?)
Deepen accounting:Details on consumption (capital/current), reallocations (capital/credit/land), and transfers (education, health, other).Develop Wealth Accounts.
Project accounts:2007 to 2100?
Widen accounts: Education and Gender
Extend accounts:Back to 1776?
Three Models
1. UN Probabilistic Population Forecast
2. Forecasting Educational Distribution
3. Forecasting Public Sector Transfers
x
UN Probabilistic Forecast:The 4 Steps of the Method
1. Select path of TFR, e0, and NMR by repeated random draws from a set of similar countries.
2. Forecast population based on standard cohort-component method.
3. Repeat steps 1 and 2; typically 1,000 or 10,000 times.
4. Calculate predictive distributions for variables of interest (population size, OADR, e0, etc).
UN Probabilistic Forecast
• Advantages of Probabilistic Forecast
– Does not use expert opinion.
– Quantifies our uncertainty about the future.
– More “realistic” projections (since allow for variability).
• Advantages of Scenario Forecast
– Uses expert opinion.– What-if scenarios.– Truly unique forecasts
without historical precedent.
UN Probabilistic Forecast
• Tells us what the future would look like based on the collective past experience of 192 UN member countries from 1950 to 2005.
• UN Probabilistic Forecasts assume: countries with similar demographic parameters are being exposed to the same set of unknown social forces which will shape their future demographic trajectories.
• Alternative Scenario Forecasts: Why this country is unique among UN member states?
UN Scenario Forecasts Are Pessimistic about Future Longevity
20 30 40 50 60 70 80 90
0.0
0.5
1.0
1.5
2.0
2.5
Gain in life expectancy during 5 year period as function of current life expectancy
Life expectancy
Ga
in in
life
exp
ect
an
cy
Based on UN World Population Prospects, the 2004 revision.
ObservedForecast
UN Scenario Forecasts Too Certain About Future Migration
UN Probabilistic Forecast for Chile
1950 2000 2050 2100
5
10
15
20
25
30
35
40
Population of Chile , 1950 to 2100
Year
Mill
ions
with historical estimates, UN scenario forecasts, and UN probabilistic forecasts.
10th
50th
90th
Historical Record UN Probabilistic Forecasts (deciles)
1950 2000 2050 2100
0.8
0.9
1.0
1.1
1.2
Year
Pro
duce
rs p
er c
onsu
mer
10th
50th
90th
Economic Support Ratio, Chile
UN Probabilistic Forecast for the United States
1950 2000 2050 2100
200
400
600
800
Population of United States of America , 1950 to 2100
Year
Mill
ions
with historical estimates, UN scenario forecasts, and UN probabilistic forecasts.
10th
50th
90th
Historical Record UN Probabilistic Forecasts (deciles)
1950 2000 2050 2100
0.8
0.9
1.0
1.1
1.2
Year
Pro
duce
rs p
er c
onsu
mer
10th
50th
90th
Economic Support Ratio, U.S.
Forecasting Educational Distributions
• The history and future of educational distributions are interesting in their own right as measures of social progress.
• In addition, these serve as important components of economic growth models and perhaps NTA accounts.
Educational Model
• Uses data from a single census to provide historical record as well as basis for forecasting the future.
• Easily replicated! Can measure the changing educational distribution of populations throughout the world over a considerable period of time.
3 Box Model
Primary PopulationBy Age and Sex
Secondary PopulationBy Age and Sex
Tertiary PopulationBy Age and Sex
Alpha
Beta
Births Deaths
20 40 60 80 100
0
20
40
60
80
100
Age
Per
cent
WomenMen
Secondary educationNone or primary education
Tertiary education
Educational Distribution From Chilean Census 2002
Progression to Secondary Level
1920 1940 1960 1980 2000
30
40
50
60
70
80
90
100
Approximate year of birth
Per
cent
WomenMenContinued Progress Scenario.No Progress Scenario.
Progression to Tertiary Level
1920 1940 1960 1980 2000
0
10
20
30
40
50
60
70
Approximate year of birth
Per
cent
WomenMenContinued Progress Scenario.No Progress Scenario.
Rapid Change in Educational Distribution of Elderly
2000 2020 2040 2060 2080 2100
0
10
20
30
40
50
60
70
Year
Per
cent
Primary
Secondary
Tertiary
Rapid Change in Educational Distribution of Work Force
2000 2020 2040 2060 2080 2100
0
10
20
30
40
50
60
70
Year
Per
cent
Primary
Secondary
Tertiary
Forecasting Public Sector Transfers
• Basic Accounting Method for Aggregate Expenditures:
B(t) = sum { b(x,e,t) * p(x,e,t) }
where b(x,e,t) = per-capita expenditures by age x, education or other characteristic e, and time t.
and p(x,e,t) = population by age, education, and time.
?
Simplifying Assumptions
Only Age and Time Matter:
B(t) = sum { b(x,e,t) * p(x,e,t) }
Simplifying Assumptions
No change in policy nor behavior.
Simple period effects on budget factors:
B(t) = sum { b(x,2005)* exp(r*(t-2005))
* p(x,e,t) }
Expenditures and Taxes by Age, Chile
0 20 40 60 80
500
1000
1500
2000
Government Expenditures and Taxes by Age: Chile 2004
Age
000s
pes
os
Expenditures Projected to Grow More Rapidly Than Taxes
Due to Population Aging
2010 2020 2030 2040 2050
1.0
1.1
1.2
1.3
1.4
1.5
1.6
1.7
Government expenditures and taxes: Chile 2004-2050
Year
Rel
ativ
e to
200
4
Taxes increase by 30% over the next 45 years due to population aging.
2010 2020 2030 2040 2050
1.00
1.05
1.10
1.15
1.20
1.25
1.30
Fiscal Tax Ratio: Chile 2004-2050
Year
Rel
ativ
e to
200
4
The Projected Fiscal Tax Ratio in Chile: 2004 to 2050
In Chile, transition to a new joint public-private system offsets projected tax increase.
2010 2020 2030 2040 2050
0.0
0.5
1.0
1.5
2.0
2.5
3.0
Year
Tril
lions
of P
esos
Pensions - Phaseout of old systemINP BudgetPA/PM Budget
Pensions
New System
Old System
Pension projection due to population aging and policy changes.
Projecting Public Transfers
• Requires expert country-specific knowledge. Like those of NTA teams!
• As we develop NTAs, we will learn more about public transfers and public/private composition of spending (education, health, etc). How this varies among countries and over time.
Future Work
• Probabilistic Population Forecasts for NTA countries.
• History and forecasts of educational distributions for NTA countries. Implications of “educational dividend.”
• Revision of a generalized public transfer model:– Assumption of fixed age shape is often wrong. Model
should allow for cohort-effects. – Major use of the model: exploring effect of policy
changes.– As we develop NTAs, we will learn more about how
budget factors differ between countries and over time.
top related