Solution Manual For Financial Statement Analysis, 13th Edition By Charles H. ...
Session 8 c discussion of majumdar ray sinha paper in session 8 c 29 august
1. A Unified Framework for the Estimation of Intra and Inter
Country Food Purchasing Power Parities with Application
to Cross Country Comparisons of Food Expenditure:
India, Indonesia and Vietnam
Authors: Amita Majumdar, Ranjan Ray and Kompal Sinha
Discussant: D.S. Prasada Rao
2. Background
The main source of cross-country price comparisons is the ICP which
provides estimates of PPPs
ICP provides one PPP for each country for GDP and components
ICP 2011 PPPs are: Rs. 15.11 per US dollar (India); Rs. 3606.57
(Indonesia) and dong 6709.19 (Vietnam)
PPPs vary across rural-urban regions
This is true. That is why PPPs must be cautiously used in poverty
computations.
PPPs vary across different regions within a country (sub-national PPPs)
PPPs from ICP are at national level using national average prices.
PPPs are not item specific and may vary across different items
ICP provides PPPs at basic headings which are the same as the
items used in this paper. In fact ICP provides more detailed level
PPPs
PPPs are calculated for fixed, country invariant baskets of items
This is not true. ICP regions use a large basket of goods and
services and countries price items that are relevant
3. Objectives
The objective of the paper is to provide a unified
framework that makes it possible to compute PPPs at:
country level
sub-national level
Item level
Aggregation methodology based on economic theoretic
approach to the computation of price index numbers
4. Methodology
Price comparisons are based on Konus price index numbers –
which are expressed as ratios of expenditures necessary to
attain a reference utility level at the prices observed in
different countries are regions.
The paper makes use of the Quadratic Almost Expenditure
system (QAIDS) – similar to that used in Feenstra, Ma and
Rao (2010).
This approach is extended in the paper by suitably modifying
the budget share equation (see next slide)
5. Methodology (contd..)
The budget share of the QAIDS system is
with
The above equation can be extended to hold for all areas (pooled)
as follows, using the item specific PPP parameters , namely, the
to express the urban prices in terms of the rural prices or,
alternatively, the PPP of the comparison country in terms of the
reference country by replacing by , where ,
with denoting sectoral/country dummy, is the OECD
equivalence scale, n being the household size.
5
6. Data Sources
• Indian data: 66th round (July, 2009-June, 2010) of India’s
National Sample Surveys (NSS) on household level
consumption expenditure (15 major states , rural and
urban).
• Indonesian data: Indonesian Social and Economic Survey
(SUSENAS) 2011, collected by the Central Statistical
Agency of the Government of Indonesia (32 provinces, rural
and urban).
• Vietnamese data: Vietnamese Household Living Standard
Surveys (VHLSS) of 2010 (3 regions, rural and urban).
6
7. Data
Data are available at the household level
Expenditure and quantity data are available
Unit values are derived and used as prices for different consumption
items like, rice, wheat, potatoes, etc.
Paper adjusts for quality
As household expenditure surveys do not have quantity data for items
other than food, analysis here is limited to “Food”.
The following 8 Food groups were constructed by
matching individual items* within the group across the 3
countries (to the extent possible). A part of the analysis is based on the
first 5 items.
Cereals & Cereal substitutes;
Milk & Milk Products;
Edible Oil;
Meat, Fish & Eggs;
Vegetables;
Fruits;
Alcohol, tobacco and intoxicants;
Beverages. ________________________________________
* Only those items, for which quantities are available, were chosen. 7
8. Food PPPs Indonesian Provinces (Jakarta = 1.00)
Province Rural Urban
Sumatra Utara 1.330 1.396
DI Yogyakarta 0.968 0.924
Bali 1.111 1.523
Kalimantan Barat 1.583 1.168
Sulawesi Utara 1.075 1.045
Papua 1.290 1.122
9. Urban PPPs Indonesian Provinces (Rural = 1.00)
by commodity groups
Province Spatial adjusted Spatial unadjusted
Cereals 1.283 1.106
Milk and milk products 0.971 1.184
Edible oils 1.312 0.965
Meat, Fish and egg 1.266 1.052
Vegetables 1.190 1.087
Fruits 1.687 2.977
10. Results: PPPs for Food from the Study and ICP
ICP
India
(Base)
PPP
Indonesia Vietnam
ICP
2100(Food)a
1 295.00 567.00
ICP 2011
(GDP)b
1 238.70 444.05
ICP 2011
(Household
Expenditure)b
1 273.25 509.18
PPPs for Food (8 items) from current study
PPPs from ICP 2011
Source: ICP Detailed Results, World Bank, 2011
Exchange rates: 1 INR = 192 IND rupiah
1 INR = 400 Viet dong
Type of
Comparison
India Indonesia Vietnam
Bilateral
Rural
1
143.26
(11.27)
1
341.82
(18.21)
- 1
2.32
(9.14)
Urban
1
141.05
(12.53)
1
371.86
(11.56)
- 1
2.94
(24.11)
Trilateral (All)
(Rural Urban combined)
1
142.63
(83.82)
357.00
(25.07)
11. Results: PPPs for Food from the Study and ICP
Type of
Comparison
India Indonesia
PPP
Vietnam
PPP
Estimated PLI
Bilateral
Rural 1 298.57
1.550
(2.57)a
1 482.64
1.207
(5.44)
- 1 1.769
0.830
(-12.07)
Urban 1 276.30
1.435
(14.12)
1 454.49
1.137
(1.88)
- 1 1.865
0.876
(-1.91)
Trilateral (All)
(Rural Urban
combined)
1 255.72 461.84
1.328, 1.155
(9.44) (6.59)
ICP
India
(Base)
PLI = PPP/ER
Indonesia Vietnam
ICP
2100(Food)a
1 1.536 1.420
ICP 2011
(GDP)b
1 1.454 1.159
ICP 2011
(Household
Expenditure)b
1 1.270 1.010
PPPs and PLIs for Food (5 items) from
current study
PLIs from ICP 2011
Source: ICP Detailed Results, World Bank, 2011
Exchange rates: 1 INR = 192.6 IND R
1 INR = 399.9 Viet dong
12. Sen’s Welfare Measure: W = μ(1-G)
Sector Sens’ W
India
μ G W
Indonesia
μ G W
Vietnam
μ G W
ICP PPPa
Rural 707.01 0.2739 513.33 671.55 0.2596 497.21 1501 0.3164 1026
Urban 820.22 0.2709 598.03 708.52 0.2723 515.59 1742 0.3231 1179
Item invariant PPP
(Trilateral PPP)
Rural 707.01 0.2739 513.33 704.33 0.2596 521.48 1044.21 0.3164 713
Urban 820.22 0.2709 598.03 540.75 0.2733 540.75 1212.18 0.3231 820
13. Main findings
The paper demonstrates the feasibility of using household expenditure data to
derive PPPs for “food” and its component commodity groups.
The methodology can be used to derive bilateral as well as multilateral PPPs
between countries
Demonstrated for India, Indonesia and Vietnam
Differences found to minimal – however they could be different if more countries are
included in the exercise.
The approach makes it possible to compute:
PPPs for different sub-regions within a country
PPPs for rural and urban areas
PPPs are important for poverty related work
14. Discussion
Differences between this approach and the ICP?
This papers makes use of unit values or prices for each commodity at the household
level.
As households are located in different geographical locations (provinces and
regions), variations in prices according to these attributes can be utilized.
ICP uses national annual average prices – these are averaged over all the
geographical and rural and urban locations.
This paper has expenditure data associated with each unit value or price observation.
This means that weights data are available at each price level.
In contrast, weights are available only at the “basic heading” level in the ICP.
ICP aggregates price data from item level to basic heading level without weights
using CPD whereas this paper uses expenditure weights for each price.
15. Comments
This an outstanding paper which is quite extensive in its scope. It provides a rigorous
approach to the computation of sub-national PPPs.
It would have been better if the actual method and estimating equations are spelt out
more clearly.
It is not clear if the rural-urban and spatial PPPs are derived from the
coefficients from dummy variables or from some sort of Konus index based on
QAIDS index number formula.
Discuss the implications of QAIDS for the “food” subgroup.
The use of term “item” is somewhat confusing.
I understand that “item” is used for a generic consumption item like “rice” or
“wheat”.
Within ICP, “rice” is a basic heading which covers 20 different varieties of rice.
In contrast, rice in this paper is a single item.
If items refer to “basic headings”, then ICP provides estimates of PPPs at the
item level.
16. Comments
The paper makes correction for quality differences when using “unit values” which are
like average prices.
However, it approach does not allow for differences in items like rice
Rice consumed in India may be of a different variety compared to Vietnam or
Indonesia.
This is also true for rice consumption in different provinces within India – in
Punjab people may consume “basmati” where as in Bengal “parboiled rice” is
consumed.
Can this method account for such differences?
Main limitation of the approach is that this can be used only for the products where
both expenditure and quantity information are available.
This means that most of the items in consumption basket cannot be included.
Can this approach be extended to include non-food items?
In particular, would it be possible to include price data for non-food items even
if they are in the form of average prices?
Ultimately, it would be good if an approach that combines the advantages of the ICP
and the use of unit values is developed.
Editor's Notes
The geometric lag has the advantage of being relatively simple. But there are many instances where it is unreasonable to assume that the first lag weight is the largest (e.g., the inflation example).
The geometric lag has the advantage of being relatively simple. But there are many instances where it is unreasonable to assume that the first lag weight is the largest (e.g., the inflation example).
The geometric lag has the advantage of being relatively simple. But there are many instances where it is unreasonable to assume that the first lag weight is the largest (e.g., the inflation example).
The geometric lag has the advantage of being relatively simple. But there are many instances where it is unreasonable to assume that the first lag weight is the largest (e.g., the inflation example).
The geometric lag has the advantage of being relatively simple. But there are many instances where it is unreasonable to assume that the first lag weight is the largest (e.g., the inflation example).
The geometric lag has the advantage of being relatively simple. But there are many instances where it is unreasonable to assume that the first lag weight is the largest (e.g., the inflation example).
The geometric lag has the advantage of being relatively simple. But there are many instances where it is unreasonable to assume that the first lag weight is the largest (e.g., the inflation example).
The geometric lag has the advantage of being relatively simple. But there are many instances where it is unreasonable to assume that the first lag weight is the largest (e.g., the inflation example).
The geometric lag has the advantage of being relatively simple. But there are many instances where it is unreasonable to assume that the first lag weight is the largest (e.g., the inflation example).
The geometric lag has the advantage of being relatively simple. But there are many instances where it is unreasonable to assume that the first lag weight is the largest (e.g., the inflation example).
The geometric lag has the advantage of being relatively simple. But there are many instances where it is unreasonable to assume that the first lag weight is the largest (e.g., the inflation example).