Comparing Consumption: inter-national and inter-temporal An-Chi Tung, 2007.1.19
Outline Graphs across economies Findings and questions patterns: right peak, left peak, twin peak, central peak… public vs private health and education Findings and questions trend? changes? why? what impact? …
Graphs of Consumption
List of Countries
Patterns of Mean Consumption: preliminary guess Peak at right end US 2003, Sweden 2003 Twin Peaks Japan 1999, Costa Rica 2004, France 2001, Korea, 2000 Peak at left with immediate drop: Taiwan 1986, 1991, 1998, 2003 with smooth drop: Thailand 1996, 2004; Indonesia 1996, 1999 Other: Increasing, flat, decreasing… Chile 1997, Uruguay 1994, India 1999, Philippines 1999?
- very high ratio for Sweden! (The US is high, too.) - due to health expenditures? mostly public?
- due to both education and health? Japan: 17…38…57 Costa Rica: 27…43…57 France: 15…43…54 Korea: 18…42…53 (Taiwan: 21..[29]..58) - due to both education and health?
- time pattern: save less over time? Peak age: 19-21 - due to education? - time pattern: save less over time?
mild peak, due to education? Thailand 2004: 21 Thailand 1996: 20 Indonesia 1996, 1999 : 23 mild peak, due to education? if so, the peak age is a little late in Indonesia (check why) - time pattern
Chile: 50
Very late! Late! Education?
Why Different? (i) time: increasing over time? Differences among age groups may have to do with (i) time: increasing over time? (ii) composition of consumption: education, health, others… (iii) private or public and financial sources: YL, asset reallocation, transfer…
Time Trend of Level and Shape
Taiwan, 1981-2003
From Gretchen Donehower : Labor Income and Consumption, 1888-2003 Dollars (US, 2000) NOTES: So you can see a few things over time: Lots of smoothing in early profiles - sparse surveys Changing shape of consumption - increasing at older ages Changing shape of labor income - narrowing years of peak earnings. How to compare across time when scale is changing? (Even after adjusting to constant dollars, economic growth has been pretty steady over the past 115 years.) USUAL NTA TRICK OF SCALING TO (UNWEIGHTED) AVERAGE LABOR INCOME VALUE FOR AGES 30-49 Age
Composition of Consumption Importance of Health and Education Size of the Public Sector
Public vs. Private Public Private education: allocated by no. of students at each level health: NHI, by individual; non-NHI, per capita the rest: per capita Private education: regression on students health: regression housing and durables: equivalence scale the rest: equivalence scale
Decomposing Consumption (normalized by mean YL age 30-49) Sweden 2003 US 2003 Japan 2003 Taiwan 2003 Indonesia 1999 Costa Rica 2003 Thailand 2004 Taiwan 1981
% of Health and Education in C (normalized by mean YL age 30-49) Sweden 2003 Japan 2004 France 2001 Uruguay 1994 Taiwan 2003 Japan 2004 Philippines 1999 Indonesia 1999
% of Public C in C (normalized by mean YL age 30-49) Sweden 2003 Japan 2004 France 2001 Uruguay 1994 Taiwan 2003 Japan 2004 Philippines 1999 Indonesia 1999
Change in Composition over Time in Taiwan (normalized to avg C of age 20-64) Taiwan, 1981 Taiwan, 2003 Larger share of public C, and larger shares of CGE and CGH for age 01-19 than in 1981 Allocating more to the youth group than to the elderly Age 0-19 20-64 65+
Taiwan, 1981 Chile,1997 Taiwan, 2003 Sweden, 2003 Taiwan, 2003 Age 0-19 20-64 65+ Age 0-19 20-64 65+ Taiwan, 2003
Issues and Questions
General Observations There are differences across economies in level, shape, composition. For example, some countries have much larger public sector than the others, and some spend more on education. For total C, there are four basic types. Longitudinally, there may be a time trend.
Some questions About trend or changes over time: - Is avg C (relative to YL) rising over time? Has this to do with GDP level or co-residence pattern? - Why does avg C drop after an early age: e.g. age 19 in Taiwan, 1998? 2. Does the increase in public consumption on health spending reduce the public consumption on education (as in the case of Taiwan)? If yes, future growth may be affected. Does the increase in public consumption on health cause a decrease in private consumption on health? …
Methodological Issues… Same definition of education and health across countries? (e.g., does “education” includes boarding expenditure and private tutoring? 2. Which methodology is best?