The development of the Indonesian service sector: a first analysis of the estimates

1951-52: Neumark

When constructing national accounts for Indonesia for 1951 and 1952 Neumark (1954) estimated commodities and articles entering into commerce by main groups of articles and commodities.

For agricultural commodities he concluded that reliable estimates were not available and therefore he chose a rather rough net margin of 10 per cent.

For trade margins on export trade Neumark took a commission of 2 per cent for all estate crops exported, except for palm oil, kernels, and Deli tobacco, since these crops are mostly exported by the producers themselves and consequently the value of the services of exporting may be retained in the value of the product instead of being added to trade.

Export crops grown by peasants were taken at a margin of 15 per cent, since such crops involve a great deal more collecting and handling than in the case of Estate crops.

For live animals and products of animals and products of animal origin an export margin of 20 per cent was allocated to trade. Other margins used by Neumark for export trade are as follows:

Table 1: Trade margins for different export commodities
Commodity	Trade margin
Estate export crops	2 %
Peasant export crops	15 %
Live animals and products of animal origin	20 %
Wood, etc.	10 %
Forest products	15 %
Other products of vegetable origin	15 %
Mining products^a	15 %

Source: Neumark (1954), p. 381.

^a: Excluding oil products, tin and bauxite which are exported by the mining companies

For other imported consumer goods, excluding imported kerosene, which is accounted for elsewhere, a margin of 40 per cent is taken. Margin on imported raw materials, and auxiliary products, which are mostly purchased in bulk by a comparatively small number of buyers, is estimated by Neumark to be 20 per cent. Other margins used by Neumark can be found in table 2.

Table 2: Trade margins used by Neumark on other commodities
Commodity	Trade margin
Durable consumption and capital goods	25 % (of which 17.5 % value added)
Locally manufactured goods	35 %
Forestry and forestry products^a	40 %
Fish	30 %
Cattle^b	±75 %
Goats and sheep	±25 %
Pigs	±75 %
Chickens/ducks^e	±35 %

Source: Neumark (1954), p. 382-384

^a: Assumed is that 40 % of charcoal and firewood produced enters into trade.

^b: Assumed is that 20 % of the cattle and buffaloes, 30 % of sheep and goats, 50 % of the pigs, 65 % of poultry and 60 % of the eggs did not enter into trade. The meat per animal was assumed to be 120 kg for cattle and buffaloes, 40 kg for pigs, 8 kg for sheep and goats.


1970s until nowadays: Estimates based Input-Output tables

With the first input-output tables constructed in 1975 the trade margins are no longer based on irregular surveys, but directly taken from this publication. So gross value added of the trade sub-sector was still estimated by using the commodity flow method. After computing the output of agriculture, mining and quarrying, manufacturing and import commodities marketed through the trade sub-sector, the results were multiplied by ratios of traded merchandise, and trade margin derived from Input-Output tables.⁷

2.3 Government

The conventional practice in national income estimation is to evaluate government services – other than those of government enterprises – in terms of expenditure made for them. Government activities can best be split into administrative and commercial activities. The former are valued at the cost of these services, that is, as equivalent to the wages and salaries paid by government administrative departments and the latter on the same basis as other productive enterprises. The government sector includes under government services the public administration. Under commercial activities railways, post and telegraph, opium production, salt production, etc. These major commercial activities are treated separately.

Government administration in the colonial period

The method employed for estimating government estimation in the colonial time is the following. From the Netherlands Indies Budgets for several benchmark years the wages paid is taken for the different departments. Assuming that the ratio of wages paid to total expenditure per department is constant we arrive at an estimate of this subsector. In table 3 figures are taken from a study by Hart (1932). We see that between 1922 and 1931 salaries made up between 29.5 per cent and 37.1 per cent of total expenditures. This corresponds with my own findings for 1907, when salaries and wages accounted for 34.5 per cent of total expenditures. It therefore seems reasonable to estimate total value added of the government sector for the colonial period as 35 per cent of total expenditures to which 5 per cent is added as approximation of depreciation, as is also done in Indonesia’s present-day national accounts.

Table 3: Salaries and wages paid by the government, 1922-1931 (in million guilders, excluding State enterprises)
	Total expenditure	Interest	Pensions	Salaries	Salaries as % of total
1922	501	73	38	148	29.5%
1923	436	79	41	145	33.3%
1924	402	85	45	143	35.6%
1925	415	83	48	152	36.6%
1926	431	83	50	160	37.1%
1927	466	84	52	169	36.3%
1928	502	98	55	176	35.1%
1929	502	86	62	182	36.3%
1930	508	81	62	185	36.4%
1931	517	85	65	192	37.1%
Source: Hart (1932), p. 313

Government administration in Indonesia’s National Accounts since independence

Figures about wages and salaries were in the period 1951-1954 directly taken from the different ministries. In the national accounts estimates of the 1960s production accounts for the central government were derived from actual routine budget expenditure statements of the Budget Directorate of the Department of Finance. Fixed capital consumption was calculated as 5 per cent of net value added, i.e. compensation of employees.

For the local government compensation of employees was calculated as the product of average annual income per employee and employment figures. Employment estimates were based on information from the Department of Interior, and on the Census of Civil Servants.

For the 1980s and 1990s the contribution of this sector to gross domestic product consisted of routine wages and salaries of central and local government employees, the wage component of the development budget, and 5 per cent depreciation. The estimation was based on realized government expenditure gathered from the Ministry of Finance and BPS.

2.4 Ownership of Dwelling

In the System of National Accounts the contribution of house property to the national income is taken as equivalent to the net rental income of dwellings. No distinction is made between owner-occupied houses and rented houses.

In the absence of any proper statistics an option to estimate gross value added is to assume that ownership of dwelling corresponds to a constant percentage of the gross value added of all other activities at both current and constant prices. This is also the method used by BPS in the 1960s, which assumed a 2 per cent contribution to GDP.

But it seems that with the available information more reliable estimates can be made. From the 1930 census we learn that there was in Java and Madura a total of 8,784,000 Indonesian dwellings. Of these 352,000 were brick dwellings, while 4,837,000 were classified as ‘other than brick dwellings with permanent roofs’. There was further a total of 48,000 European dwellings (mainly brick), 121,000 Chinese and 11,000 other dwellings. Figures for the rest of Indonesia are not available but can be estimated from this.

Besides knowing the number of dwellings per population group in Java and Madura, the 1930 census also tells us how many people from the different population groups were living in Java and Madura. From this we can calculate for the different population groups how many people on average lived in one house. If we assume that this ratio was equal in the Outer Islands we can derive the number of dwellings in the Outer Islands by multiplying this ratio by the number of people of the different population groups living in the Outer Islands. This information is also taken from the 1930 census. In this way we arrive at an estimate of the total number of dwellings in Indonesia in 1930. The results are shown in table 4.

Table 4: Number of dwellings in Indonesia, 1930
	Indonesian	European	Chinese	Other	Total
Number of dwellings in Java & Madura	8,784,000	48,000	121,000	11,000	8,964,000
Population in Java & Madura	40,891,093	192,571	582,431	52,269	41,718,364
Ratio population/dwelling in Java & Madura	4.7	4.0	4.8	4.8	4.7
Population in Outer Islands	18,246,974	47,846	650,783	63,266	19,008,869
Estimated number of dwellings in Outer Islands	3,919,715	11,926	135,200	13,314	4,084,424
Total number of dwellings in Indonesia	12,703,715	59,926	256,200	24,314	13,048,424
Sources: Own calculations based on Population census 1930

The available data do not allow constructing time series for prices of different categories of houses. Therefore we take 1930 as a base year. The population census enables us to divide the houses into three categories: 1) Upper-class: European brick dwellings (48.000), 2) Middle-class: Indonesian brick dwellings (352.000) and Chinese dwellings (121.000), and 3) Cheap houses: others. Multiplying these figures with an average rent of f 15 for cheap houses, f 50 for middle class houses and f 100 for upper class houses we arrive at an estimate of f 154.930.000 per month for Java and Madura.⁸ Assuming that the same share of the Indonesians in the Outer Islands live in brick dwellings (i.e. 4 per cent), we can extrapolate this figure for Indonesia as a whole. This results in a total annual rent paid of fl. 2.725.991.112. This figure however also includes maintenance and other costs which are not considered value added. Therefore we have to find a value added/output ratio to arrive at an estimate of the value added of the housing sector in 1930. This will be taken from the Input-output tables.