Application of Cutoff Sampling in Enterprise Surveys for Financial Statistics the Case of Foreign Di

1
Application of Cut-off Sampling in Enterprise
Surveys for Financial Statistics the Case of
Foreign Direct Investments

• Jorma HilpinenBank of Finland

2
Resources, Efficiency, Quality

• Declining resources
• Changing information requirements and economic
  structures
• Low quality statistics not tolerated
• Dominance and extreme skewness problems plague
  enterprise sector data in small open economies

3
Resources, Efficiency, Quality

• Efficiency demands override statistical orthodoxy
• Streamlined processes for efficiency
• Registers, frames and respondent selection are
  prominent quality factors
• Practical and simple depiction of skewness
• Descriptive statistics of auxiliary variables
• Statistica box-plot an easy-to-use tool
• Use of balance sheets instead of survey data for
  small enterprises

4
Survey year
5
Stratification of the frame

• Vertical axis cumulative coverage of a target
  variable, per cent (0 to 100)Horizontal axis
  number of enterprises (0 to n)xyzw100a
  very large enterprises, b medium- size
  enterprises, csmaller enterprises surveyed d
  not surveyedabcdframeabcsurveyed
  enterprises

6
Stratification of the frame
7
Stratified auxiliary variable data (mean, median
and standard deviation) , million euros,
coefficient of variation (cv)

• Data consisting of enterprises with ID match
  between FDI and balance sheet data.
• Zero and negative observations excluded from each
  group
• auxiliary variables equity and
  liabilities to shares in receivables
  from
• reserves affiliates
  affiliates affiliates
• target variables1 A B C D

stratum Ten largest n 10
10 10 10
Mean
641.4 313.6 1412.5 281.5 median 406.0
204.2 1189.0 138.0 std 767.8
255.9 1120.7 342.9 coefficient of variation
1.2 0.8
0.8 1.2 (cv
std/mean) Medium size n 109
34 17 23 mean
55.4 47.8 281.9 96.1median 36.0
42.4 183.0 41.2 Std 52.8
32.9 264.6 202.5 Cv 0.9 0.7 0.9 2.1
1 Target variables A. FDI in Finland equity,
B FDI in Finland other capital, C FDI abroad
equity, D FDI abroad other capital 2 Outliers
and extremes excluded according to the Statistica
criteria. FDI abroad In addition, the
enterprises with exceptionally low FDI in
relation to the auxiliary variable were excluded
8
Stratified auxiliary variable data (mean, median
and standard deviation) , million euros,
coefficient of variation (cv)

• Data consisting of enterprises with ID match
  between FDI and balance sheet data.
• Zero and negative observations excluded from each
  group
• auxiliary variables equity and
  liabilities to shares in receivables
  from
• reserves affiliates
  affiliates affiliates
• target variables1 A B C D

Small enterprises surveyed n
195 100 132
64 mean 11.9 13.4
70.1 55.7 median 8.1
7.8 16.1
20.3 std 24.6 20.2
220.0 107.3 cv
2.1 1.5 2.3
1.9 Small
enterprises, surveyed, 'good leverage
points'2 n 173
86 111 27 Mean 8.3
8.9 17.9
5.5 Median 7.4 6.5 8.8 3.1 std 5.5
8.0 19.8
5.0 cv 0.7 0.9
1.1 0.9 Small
enterprises, not surveyed n
571 533 448 532 Mean 3.1
2.9 12.3 10.5 Median 1.4
0.7 1.0 0.9 Std 6.2
10.0 48.9 37.9 Cv 2.0 3.4 4.0 3.6
1 Target variables A. FDI in Finland equity,
B FDI in Finland other capital, C FDI abroad
equity, D FDI abroad other capital 2 Outliers
and extremes excluded according to the Statistica
criteria. FDI abroad In addition, the
enterprises with exceptionally low FDI in
relation to the auxiliary variable were excluded
9
Surveyed small enterprises, target variables
10
Surveyed small enterprises, target variables
11
Surveyed small enterprises, auxiliary variables
12
Surveyed small enterprises, auxiliary variables
13
Regression example
14
Regression example