Nonfinancial Corporations

12. Data for the current, capital, and financial accounts, together with their balance sheets, of nonfinancial corporations (NFCs) are drawn from a variety of sources. For private NFCs, the BPS conducts an annual survey of unlisted enterprises, and the BPS obtains data on listed enterprises from information placed on the Indonesian Stock Exchange (IDX) for annual and quarterly data. For public NFCs owned by the central government, the Ministry of State Owned Enterprises (MOSOE) provides data on these accounts to BPS from annual and quarterly reports it receives from these enterprises, both listed and unlisted. These data are also compared with, and supplemented by, annual and quarterly account information for those public enterprises listed on the IDX.

13. For listed private corporations, the data are comprehensive (both quarterly and annually), but the data drawn from the IDX are only available in PDF format. As a result, BPS has to enter the information manually, making for a very slow, inefficient process, subject to input errors. It would be much more efficient, and much faster, if the data could be provided electronically. The mission recommended that OJK, who obtain these data from the private listed enterprises, provide BPS with the data, on an individual enterprise basis, in a machine-readable format (Excel or another software package) that allows for the data to be easily entered and aggregated much more easily and efficiently.

14. While the annual survey of private NFCs provides BPS with the data required to meet the needs for the annual sector accounts, quarterly data for unlisted private NFCs pose BPS with problems, as there are no direct sources. For the current and capital accounts, the BPS uses ratios between the annual data of listed and unlisted NFCs, but for equity of unlisted NFCs, the BPS has no current data sources. In light of this, the previous mission had recommended that the BPS run a regression between the data for annual equity listed and unlisted enterprises to determine whether there is a correlation that would allow the latter to serve as an indicator for the former for quarterly data.

15. The BPS had conducted a regression for the years 2011 to 2014 (the only years for which data are available) and found a very high correlation and statistical significance. In light of this, the mission recommended that the BPS conduct another regression analysis of changes in positions in equity of unlisted against changes in equity positions of listed enterprises to determine whether the latter could serve as an indicator for transactions of the former. The mission also suggested that, data permitting, the BPS do the same exercise for “other accounts receivable/payable.” The mission recommended that this regression be conducted before the next mission.

16. For the sector accounts’ needs for quarterly data for public NFCs owned by the central government, the MOSOE obtains data for the 18 largest enterprises, which are then grossed up to the universe estimate. Those 18 enterprises account for a very large proportion of the total, producing high quality estimates. The MOSOE provides BPS with these data on a timely basis.

17. No data are available for public NFCs owned by other levels of government. However, BPS and MOSOE do not consider them significant.

18. The previous mission had proposed that BPS and MOSOE sign a memorandum of understanding (MOU) to ensure that all the data, especially quarterly, that BPS needs to construct the sector accounts (quarterly and annually) are reported on a timely basis. Since that mission, the MOSOE has been providing the data required on a timely basis. Given that, it may not be necessary to sign an MOU. This matter is to be discussed between BPS and the MOSOE over the next few weeks and a decision will be taken whether or not to proceed with an MOU.

Recommended Actions:

• BPS conduct a regression analysis of changes in positions in equity of unlisted against listed enterprises to determine whether the latter could serve as an indicator for quarterly transactions of the former. (June 2018)

• Further, data permitting, BPS do the same exercise for “other accounts receivable/payable.” (June 2018)

• OJK provide BPS with data of private listed enterprises, on an individual enterprise basis, in a machine-readable format (Excel or another software package that allows for the data to be easily entered and aggregated much more easily and efficiently). (March 2018 and ongoing thereafter)

• BPS and the MOSOE sign an MOU, if necessary, for the delivery by the latter to the former of the data needs for the sector accounts and balances on an ongoing basis, both quarterly and annual. (December 2018)

General Government

19. For central government, the data are comprehensive on a monthly basis. The Ministry of Finance (MOF) collects data on the current, capital, and financial accounts monthly and reports these data quarterly to BPS, on the basis of Government Finance Statistics (GFS). These data are largely sufficient to meet BPS’ needs in compiling the quarterly sector account for central government. BPS makes some adjustments to the data to reflect general government data for national accounts’ requirements. For the balance sheet, the MOF compiles quarterly financial assets and liabilities, using the most recent annual and semi-annual positions data and adding transactions. The MOF adjusts the data, where necessary, for changes in exchange rates but not for price changes. However, because the Government of Indonesia operates on a partial accrual basis, its pensions to its (current and former) employees are treated on a pay-as-you-go basis. Accordingly, the MOF does not recognize the pension liability. There are no current plans to change this treatment. The previous mission had recommended that, as the sector accounts and balance sheets are an experimental dataset, at least for the life of the project, this shortcoming would have to be accepted, without trying to make any ad hoc adjustments.

20. The previous mission had also recommended that a reconciliation table between the two datasets (GFS and national accounts) be prepared, showing where the differences lie in balancing the accounts. This reconciliation table has been constructed.

21. For the provinces and municipalities, data are obtained by the MOF. Annual data are sufficient for the purposes of the sector accounts, but the quarterly data are less detailed and are provided only semi-annually, with a three-month lag. Consequently, the MOF uses the central government’s patterns of expenditure, based on the provinces’ and municipalities’ budgets, to estimate the provinces’ and municipalities’ outlays, by quarter.

22. The third level of government – villages – is not recognized by the MOF as part of general government. Consequently, no data are reported. The previous mission had suggested that, as a preliminary measure, the cash transfers from central government be used as a proxy for villages’ revenue. These transfers make up a major part of villages’ revenue (probably in excess of 90 percent). However, these transfers do not show how the funds are spent. Such detail (and any income that the villages obtain from non-government transfers) are only available annually, with a significant (two year) lag. The previous mission recommended using the cash transfers as a basis for deriving current quarterly data, together with the ratios of other income and uses, from the annual data and revising the data when more recent years become available. This work has not yet been started. The mission recommended that this work be undertaken as soon as possible.

23. The treatment of a further level of government, the social security fund, is yet to be resolved. This fund was set up in 2014 and all the parties involved have agreed that it should be treated as part of general government. The exact manner of its treatment, however, has not yet been agreed. The previous mission had recommended that the treatment of the fund be taken up by the working and/or technical groups and resolved by the time of the current mission. The mission emphasized that a decision needs to be made soon.

Recommended Actions:

• MOF and BPS begin estimating quarterly revenue and expenditure for villages, using the cash transfers from senior levels of government, together with the ratios of other income and uses from the annual data. (December 2018)

• BPS and MOF, and any other agencies involved, determine the treatment of the social security fund and develop quarterly estimates from the first quarter of 2015 on. (This treatment is to be resolved as soon as possible)

Financial Corporations: The Central Bank and Other Deposit-taking Corporations

24. BI has responsibility for compiling the current, capital and financial accounts, together with their balance sheet data, for the central bank and for the other deposit-taking corporations (ODTCs). These data are obtained monthly and reported, or will be reported, to BPS quarterly. There do not appear to be any major issues with these datasets.

Financial Corporations, Other than the Central Bank and Other Deposit-taking Corporations

25. The current, capital, and financial accounts, together with the balance sheet data, for the other financial corporations not collected by BI are obtained by the Financial Services Authority (OJK) monthly and reported to BPS. The mission found these data to be of high quality but some areas need further development.

26. One such area is the treatment of mutual funds. These funds are currently maintained as segregated funds of securities companies that operate them, and OJK includes them with the securities companies’ data. As these funds have full sets of accounts, the mission recommended that they should be separated from their parent companies, and classified either as money market funds or nonmarket funds, depending on the nature of their assets and activities. This change in treatment was agreed to by BPS and OJK.

27. Another area where further work is required is an estimation technique for approximating the market value of equity liability of financial corporations, other than for the central bank and ODTCs. At present, the own funds at book value approach is used but OJK and BPS are planning to explore other possibilities to see whether a better methodology can be developed.

28. OJK’s data on mutual funds also indicate some of their shares are held by nonresidents. BI does not currently include these data in the balance of payments or the international investment position (IIP). This issue is discussed further below under Rest of the World.

Recommended Actions:

• OJK, working with BPS, treat mutual funds as separate institutional units from their parent securities companies, and classify them either as money market funds or nonmarket funds, depending on the nature of their assets and activities, and provide BPS with quarterly data from the first quarter of 2015 on. (June 2018 and ongoing thereafter)

• OJK and BPS develop an improved methodology for estimating the market value of the equity positions of unlisted financial corporations other than ODTCs. The work should be undertaken in with BI, which will be involved in a similar exercise (with BPS) to improve the methodology for the measurement of nonresident ownership of the equity positions of unlisted enterprises in Indonesia, notably for direct investment enterprises. (June 2018)

Rest of the World

29. BI has the responsibility to compile and publish quarterly Indonesia’s balance of payments and IIP. One area of concern with these data is the data on the export and import of goods and services. In principle, exports and imports of goods and services in the balance of payments data should be identical to the same series in the expenditure side of gross domestic product (GDP), published by BPS. However, these two datasets are not identical, and there have been differences for many years. BI and BPS have agreed to use the data from the supply and use table (SUT) for 2010 as the final figures for that year. However, as there are no SUTs for the succeeding years, differences between the two datasets remain.

30. This issue highlights one of the important roles of the sector accounts: statistical consistency. Within the sector accounts integrated framework, inconsistent datasets are confronted and need to be resolved. Otherwise, either the matrix will not balance or there are differences with other data that should be the same. In this instance, if there is an inconsistency between the balance of payments and the expenditure side of GDP and the latter estimate is used (to be consistent with GDP, which is what BPS considers ideal to do, and would be explored further by BI), the difference will result in a different external balance of the rest of the world in the sector accounts from the current account balance in the balance of payments. BI and BPS indicated that to prevent this, they wish to make an equal and opposite adjustment to the income account (on reinvested earnings) of direct investors. The mission argued against this course of action: to rectify one error by creating another tends to defeat the point of statistical integration. The mission emphasized the need for complete coherence between the two datasets. BPS and BI recognized the importance of the issue and plan to address the matter very soon. They noted, however, that, because it is a complex issue, it may take some time to resolve. Therefore, more coordinated efforts to improve the coherence of those two datasets should be pursued. The mission recommended that BPS should compile the SUT annually to avoid a repetition of this issue as well as to provide a benchmark for all goods and services in GDP.

31. As noted, OJK’s data show that nonresidents have owned shares in mutual funds for several years, again, highlighting how the sector accounts’ integrated framework can highlight data shortcomings. To incorporate these data into the rest of world, and to ensure that the rest of the world’s data are consistent with the balance of payments and IIP, BI agreed with the mission that the changes will be made from 2015 onwards.

32. Further work that is required is an estimation technique for approximating the market value of equity liability of unlisted enterprises with nonresident ownership, notably, direct investment enterprises. At present, the own funds at book value approach is used but BI and BPS will explore other possibilities to see whether a better methodology can be developed. Including OJK (involved in a similar exercise (with BPS) for the measurement of the value of equity positions in unlisted financial corporations other than the central banks and ODTCs, would assist the outcome.

Recommended Actions:

• BPS and BI determine the causes of, and eliminate, the differences in goods and services imports and exports between the balance of payments and the expenditure side of GDP from 2011 on. (September 2018)

• Using data from OJK, BI include in the quarterly balance of payments and the IIP nonresident holdings of shares in resident mutual funds, from 2015 on. (June 2018)

• BI and BPS develop an improved methodology for estimating the market value of unlisted corporations’ equity positions with nonresident ownership, notably direct investment enterprises. The work should be undertaken in conjunction with OJK (with BPS), which will be undertaking a similar exercise for the valuation of the equity positions of unlisted financial corporations (other than the central bank and ODTCs), as noted above. (June 2018)

• BPS begin compiling the SUT annually. (December 2018)

Households

33. For households, the principal recommendation of the previous mission was to explore how to obtain estimates of quarterly production data. As there are currently no such data, that mission had recommended that BPS run a regression between households’, listed and unlisted NFCs’ annual production data for those industries – agriculture, manufacturing, and wholesale and retail trade – that account for about three quarters of household production. The purpose of the regression analysis was to see whether there is a correlation between these data for these three groups that would allow NFCs’ production data to serve as an indicator series for households’ production.

34. BPS conducted the regression analysis for 2010 to 2015 (the years for which such data are available) and found a high correlation and statistical significance between unlisted NFCs’ production data and those for households. However, there are no quarterly production data for unlisted enterprises, which means they cannot serve as an indicator for households’ quarterly production. The correlation between households’ annual production data and those for listed NFCs has a lower correlation, but still reasonably good for manufacturing and trade (87 percent and 75 percent, respectively) but the data failed the statistical significance test. Even so, given the high correlation and the absence of alternatives, the mission recommended that listed NFCs’ production data be used as an indicator for the quarterly distribution of household production for manufacturing and trade. For agriculture, the correlation is weak (45 percent). Even so, as the industry is dominated by household enterprises, the mission recommended that the quarterly data for agricultural production be used. For all other industries, the mission recommended that the average of all NFCs’ production data (other than agriculture, manufacturing and trade) be used as the quarterly indicator before the quarterly household survey conducted in 2018.

Recommended Actions:

• BPS estimate the quarterly distribution of household production for manufacturing and trade, using listed NFCs’ production data an indicator. (June 2018)

• BPS estimate the quarterly distribution of agricultural production by households from total quarterly data for agricultural production for the whole economy. (June 2018)

• BPS estimate the quarterly distribution of household production for all other industries, using the average of all NFCs’ production data (other than agriculture, manufacturing and trade) as the quarterly indicator. (June 2018)

Nonprofit Organizations Serving Households

35. As noted in the previous mission report, data for nonprofit organizations serving households (NPISHs) are reasonably complete, both quarterly and annually. The annual survey of NPISHs indicates that non-market output accounts for virtually all NPISH output (ninety-six percent) and transfers, principally from households, corporations and government, account for ninety-four percent of NPISH revenue. BPS uses the ratio of the breakdown of expenses in the annual survey to apportion these expenses quarterly. Data are cross-checked against other information, such as compensation of employees in the Labor Force Survey, and transfers from regional governments from their latter’s budgets.

36. Three areas require further consideration: how to take into account the election cycle, religious holidays, and gross fixed capital formation (GFCF), in the quarterly accounts.

37. In 2018, there are elections at the provincial and municipal level of government, and in 2019, there will be elections for the president and People’s Consultative Assembly. Political parties are classified as NPISHs, so their expenditures can have a significant impact on the overall output of NPISHs in those two years. The mission discussed with BPS staff how to allocate the expenditure by quarter when the information is only available annually. It was agreed that the expenditure should be based on the Boundary of Funding Campaign that the political parties agree with the Electoral Commission before the elections are held, and to which boundary the parties must adhere. The mission recommended that the allocation to quarter should be pro-rated to the number of days the election campaign is conducted in each quarter. Campaigning is usually for sixty days.

38. There are five major religious holidays in Indonesia, one of which has fixed a date (Christmas) and four are not fixed (Ramadhan, Idul Fitri, Idul Adha, and Easter). The quarterly survey provides sufficient detail to identify the religious bodies’ expenditure patterns by quarter, but not the level.

39. There are no sources to estimate NPISH GFCF quarterly. BPS has explored using the residual approach, that is, from the GFCF total for the whole economy deduct those sectors’ GFCF where it can be identified and make certain adjustments where data are not sufficiently robust, to leave the residual for NPISHs. The mission expressed reservations about residual derivation, given that all the errors and omissions in the other datasets would be included in the NPISH estimate. Moreover, as the value of NPISH GFCF is small (usually less than 5percent of the total) negative GFCF could result.¹

40. Therefore, the mission recommended that BPS also explore using the ratio of NPISH annual GFCF to total annual GFCF for the economy and apply that ratio to total GFCF for the economy as a whole to derive an estimate of quarterly NPISH GFCF. The mission noted that, because the most recent year may not necessarily be representative of NPISH GFCF, BPS use the average of the ratio of NPISH GFCF to the total economy’s GFCF for the previous five years. When new annual data are available, the original estimates should be revised. The mission suggested that further work on deriving estimates of GFCF by sector be undertaken in time for discussion with the next mission.² In particular, the comparisons of the two approaches – the residual calculation and the pro-rating for NPISH GFCF – be undertaken before the next mission so that the results can be reviewed then.

Recommended Actions:

• BPS prorate political parties’ expenditure by quarter, based on the number of days the election campaign is conducted in each quarter. (June 2018)

• BPS explore two approaches to the estimation of quarterly NPISH GFC, that is, residually, and using the ratio of NPISH annual GFCF to the total annual GFCF for the economy and apply that ratio to total GFCF for the economy, and BPS use the average of the ratio of NPISH GFCF to the total economy’s GFCF of the previous five years. (June 2018)