Creating a farm register the statistical units

The administrative units in taxation data, the administrative units in IACS applications and the statistical units in the Farm and Business Registers must be linked so that different sources can be integrated. The methodology for this linkage requires that different kinds of derived statistical units are created. The following administrative and statistical units should be considered here:

• Agricultural holding (AH): A single unit which has single management and which undertakes agricultural activities either as its primary or secondary activity. The target population of the Farm Register consists of this kind of units.

• Legal unit (LeU): A unit that are responsible for reporting to tax authorities or other authorities. Almost all administrative data we use for statistical purposes come from such legal units. Each legal unit has a unique identity number used for taxation purposes.

• For the Business Register a number of statistical units are created: enterprise units (EU), local units (LU), kind of activity units (KAU) and local kind of activity units (LKAU).

The structure and relations between different legal units can often be complex. This must be considered when we use administrative data. Table 2.3, with data from the Structural Business Survey, illustrates that data from related legal units sometimes must be aggregated into data describing an enterprise unit.

The relations between different kinds of units in the Business Register are illustrated in Figure 2.7. Agricultural holdings almost correspond to local kind of activity units the agricultural part of a local unit.

Table 2.3 One enterprise unit consisting of four legal units.

Turnover, SEK millions

Wage sum, SEK millions

Source 1

Source 2

Source 1

Source 3

EU 1

LeU 1

0.1

EU 1

LeU 2

8.6

1.3

EU 1

LeU 3

0.2

EU 1

LeU 4

0.6

1.2

EU 1

Sum:

8.6

8.8

1.3

Enterprise unit, EU

Local unit, LU

Kind of activity unit, KAU

£ Local kind of activity unit, LKAU

Agricultural holding

Figure 2.7 Statistical units in the Business Register and Farm Register.

If we wish to integrate data from two administrative sources (e.g. turnover from the VAT Register and wheat area from the IACS Register), we must consider the relations between different kinds of units and have access to a Business Register and a Farm Register where these relations are stored. The example in Figure 2.8 illustrates how the two administrative registers should be matched and also how the statistical estimates depend on the matching method.

The first record in Figure 2.8 is perhaps the most common case - there is a one-to-one relation between all kinds of units.

The record with LeU 2 and LeU 3 could be one holding where husband and wife both report income as self-employed but only one of them applies for subsidies. They are then registered as two different legal units by the tax authorities.

The record with LeU 4 and LeU 5 could be one holding where husband and wife both report income as self-employed and both apply for subsidies for different parts of the agricultural activities.

The record with LeU 6 describes a case where one enterprise has two local units and two holdings. The enterprise sends in one VAT report but applications for subsidies are sent in for each holding by different holders.

The last record with LeU 8 describes a case with one local unit and one holding, but agriculture is the secondary activity. The local unit is the divided into two local kind of activity units. As a rule, we have information in our Swedish Business Register describing the proportions of each activity, here 60% forestry and 40% agriculture. In Table 2.4, we illustrate how data from the two registers should be matched and how the estimates are influenced by the matching method.

The correct way of integrating these two administrative registers is shown in columns 1 and 2. For each holding we add the values that belong to it, e.g. 75 = 45 + 30 for AH 2.

For holdings AH 4 and AH 5 we have only one common VAT report. The turnover value of 50 can be divided between the two holdings by a model that describes turnover as proportional to some measure of the production on each holding.

For holding AH 6 we have one VAT report for the forestry and agricultural parts together. With the information that the agricultural proportion is 40% we estimate the agricultural part of the turnover as 0.40 ■ 100.

Columns 3-8 in Table 2.4 illustrate the consequences of matching the two administrative registers directly with the legal unit identities. Due to the fact that we do not use the correct statistical units we get mismatch in two cases. This gives us two missing o

VAT Register Business Register

Legal Turn- Legal unit over unit

Enterprise Local unit unit

Local Kind of Activity unit

Farm Register IACS Register

Agricultural holding

Legal Wheat unit area

LeU 1

120

LeU 1

LeU 2

45

LeU 2

LeU 3

30

LeU 3

LeU 4

150

LeU 4

LeU 5

80

EU 2

EU 3

LU 1

LU 2

LKAU 1 NACE 01: 100%

LKAU2 NACE 01: 100%

LKAU 3 NACE 01: 100%

AH 1

AH 2

AH 3

LeU 1

LeU 2

LeU 4

LeU 5

/

LeU 6

50

LeU 6

-

EU 4

< ,

N I

LeU 8

100

LeU 8

-

EU 5

LU 5

LU 6

LU 4

LU 5

LU 6

LKAU 4 NACE 01: 100%

LKAU 5 NACE 01: 100%

AH 4

AH 5

AH 6

LeU 6

LeU 7

LeU 8 100

Figure 2.8 Statistical units in the VAT, Business, Farm and IACS registers.

Table 2.4 Integrating the VAT Register and the I ACS.

Model for imputation Imputed values

Table 2.4 Integrating the VAT Register and the I ACS.

Model for imputation Imputed values

VAT

IACS

VAT

IACS

VAT

IACS

VAT

IACS

Turnover

Wheat area

Turnover

Wheat area

Turnover

Wheat area

Turnover

Wheat area

(1)

(2)

(3)

(4)

(5)

(6)

(7)

(8)

AH1

120

60

LeUl

120

60

120

60

120.0

60.0

AH2

75

20

LeU2

45

20

45

20

45.0

20.0

AH3

230

90

LeU3

30

No hit

No hit

30.0

16.0

AH4

50 ■ p

20

LeU4

150

60

150

60

15.0

60.0

AH5

50-(1 -p)

10

LeU5

80

30

80

30

80.0

30.0

AH6

0.40 • 100

100

LeU6

50

20

50

20

50.0

20.0

LeU7

No hit

10

No hit

18.8

10.0

LeU8

100

100

100

100

100.0

100.0

Sum

515

300

575

300

545

290

593.8

pc ryi values that can be handled in different ways. If we use the six records in columns 5 and 6 we can use the relation between turnover and wheat area (1.88) and the corresponding relation between wheat area and turnover (0.53) to impute values in columns 7 and 8. The errors created by wrong matching method are then 593.8 — 515 for the turnover estimate and 316 — 300 for the wheat area estimate.

In a calendar year register, which is discussed in the previous section, units should be combined into complex units that follow a holding during the calendar year. If, for example, holder A takes over a holding from holder B during the year, two legal units must be combined into one unit. Administrative data from these two legal units should be combined to describe the new enterprise unit.

0 0

Post a comment