Page 1
Comparison of Simulation Methods Using Historical Data in the
U.S. International Price Program
M.J. Cho, T-C. Chen, P.A. Bobbitt, J.A. Himelein,
S.P. Paben, L.R. Ernst, and J.L. Eltinge
U.S. Bureau of Labor Statistics
[email protected]
ICES III Session 14 – June 19, 2007
Disclaimer: The views expressed in this paper are those of the authors and do not necessarily reflect the policies of the BLS
Page 2
2
OutlineI. U.S. International Price Program (IPP)
II. Bootstrap Variance Estimator
III. Adequacy of Approximations in Simulation Studies
IV. Simulation Methods Considered
V. Numerical Results
VI. Conclusions
Page 3
3
I. U.S. International Price Program(IPP)
A. IPP is the nation’s primary source of information on price trends in the
international trade of the U.S. economyExport and Import Price Indexes
B. IPP involves population structure, sample design and estimation methodswith a high degree of complexity
Page 4
4
C. IPP Index Aggregation Tree
Upper Level Strata
Lower Level Strata
Classification Groups
Weight Groups (Company|CG)
Items
Page 5
5
1. Sample frame provided by U.S. Customs Border Protection, and divided into two biannual panels
2. Stratified multistage sampling within panels
a. Within a broad product category (stratum), select establishments proportional to trading dollar value
b. Within establishment, select detailed product categories (CGs) using systematic PPS
c. Within Company|CG, select items
D. IPP Sampling Design (Imported Goods)
Page 6
6
E. IPP Index Formula (modified Laspeyres index)
11
tparent
childchild
tchild
childchild
tchild
tparent LTR
WLTR
WLTRLTR
1 ttt LTRSTRLTR
Page 7
7
F. Weights
1.Item weight is weight group’s weight divided by number of items
2.Weight group’s weight is a trade dollar value divided by selection probability
3.CG weight is based on trade weights from the Census Bureau
4.Stratum weight is aggregation of CG weights
Page 8
8
1. Draw PSUs (establishments) with replacement
from the sampled PSUs in each stratum
2. Define bootstrap weights:
II. Bootstrap Variance Estimator
A. Based on Rao, Wu and Yue (1992)
selected is PSU that the timesofnumber theis
and , stratum of PSUin item for the
weightssample original is where
1
*
**
in
hik
w
nn
nww
hi
hik
hih
hhikhik
1hn
hn
Page 9
9
2
1
1 ˆ ˆˆ
where
ˆ is a bootstrap STR estimator and
ˆ is the STR estimator from the original sample
B
BT b fullb
b
full
VB
3. Calculate the price STR using bootstrap weights
4. Repeat steps 1-3 B times
5. Compute the bootstrap variance estimator:
Page 10
10
B. Properties of Under Realistic Approximations to IPP Population Structure,
Sample Design and Estimation Methods
1. Over R replications of the “simulation design” compute
1
1
21
1
ˆ ˆˆ ˆ , ,
ˆ ˆ ˆ 1 -
R
BT r BT rr
R
BT BTBT rr
V R V
V V R V V
BTV̂
Page 11
11
2. Evaluation Criteria
1 ˆˆ ˆ 11
1
21
BT
R
rr VR
21 ˆ 2 ˆ ~
BTBT VVV
Relative Bias =
Degrees of Freedom =
Page 12
12
III. Adequacy of Approximations in Simulation Studies
A. Notation
X Features of population and design( True, Approximation )
ˆ Estimator, with cdf ),ˆ( XF
),ˆ( XFQ Specific functional ofe.g. bias or MSE
),ˆ( XF
TX AX
Page 13
B. Adequacy of Simulation & Approximation
)},ˆ{{ˆAXFQ for true
1. Components to include in a. Population structure b. Sampling steps c. Nonresponsed. Weighting e. Variance estimation
2. Approximations for each component in (1)
ˆ{ { , )}TQ F X ?
AX
Page 14
14
*
2
*
ˆ ˆˆ{ { , )} { { , )}
ˆ{ { , )}
ˆ1 { { , )}
2
(Replication Error)
where
(1 ) for 0,1
T
A T
A T
X X
A T A T
X X
A T
Q F X Q F X
Q F XX X
X
Q F XX X X X
X X
X X X
1. Taylor expansion: under conditions
Page 15
15
2. Four Cases
Case 1: Small first derivatives, small
Case 2: Small first derivatives, large
Case 3: Large first derivative
Case 4: Large second derivative
TA XX
TA XX
Page 17
17
IV. Simulation Methods Considered
A. Use a Single Method for Selection of Sample Units and Related Weights:
Same sample units and weights for each month
B. Three Methods to Construct Population of Item-Level STRs
1. Resampling Method2. Fixed-One-Rate Method3. CDF Interpolation Method
Page 18
Stratum Stratum Description
P07 Edible vegetables, roots, and tubers
P08 Edible fruit and nuts; peel of citrus fruit or melons
P09 Coffee, tea, mate and spices
P22 Beverages, spirits, and vinegar
P61 Articles of apparel and clothing accessories
P74 Copper and articles thereof
P90 Optical, photographic, measuring and medical instruments
V. Numerical ResultsA. Selected Strata
Page 19
Item STR1 of Design Strata
Page 20
Original Variance of STR (P90)
Page 21
Relative Bias of STR (P90)
Page 22
Degrees of Freedom of STR (P90)
Page 23
23
VI. Conclusions
A. For Complex Establishment Surveys Like IPP, Simulation-Based Evaluations of Est Properties Require Consideration of
1. Approximations to the true population, design and estimation procedures
2. Adequacy of the resulting approximations
to the true properties
AX
ˆ{ ( , )}AQ F Xˆ{ ( , )}TQ F X
Page 24
24
B. Future Work
1. For IPP: Other features of the pop and design
e.g. Independence of STRs from sample units
2. Consider generalized variance estimator to improve stability
3. Explore the surface defined by
in “neighborhoods” of the true
TX
ˆ( , )Q F X