The sustainability of Portuguese fiscal policy from a historical perspective

Abstract

This paper analyses the sustainability of Portuguese public finances, making use of a long dataset with more than a full century of observations. The use of such a long dataset is appropriate because both the unit root and co-integration tests on which the sustainability testing procedure is based require a long period of data. We find considerable evidence in favor of sustainability for the 1903–2003 period which is not, however, maintained for the more recent 1975–2003 period, as it is characterized by the largest GDP deficit ratios of our sample. This latter period appears to signal a shift to an unsustainable path in Portuguese fiscal policy. Hence, our results suggest that fiscal consolidation efforts must, in fact, be continued in Portugal.

Statistical appendix

In Portugal, the relevant economic years for fiscal data have not always coincided with calendar years. The economic years from 1834–1835 to 1933–1934 began on 1 July of each civil year and ended on 30 June of the following civil year. The economic years from 1936 on coincided with the civil years. For the purposes of transition, the economic year of 1934–1935 began on 1 January 1934 and ended on 31 December 1935, under the terms of the same decree-law. As GDP data is only available on a calendar basis, there is inevitably a timing problem in defining fiscal variables as GDP ratios until 1936. Because fiscal decisions were taken on an annual basis (expressed in economic years), it makes no economic sense to adjust the economic years to calendar years. In order to enable a smooth transition in 1936, we deflate the economic year observation t to t+1 with the GDP of period t+1. For example the data on the fiscal year 1934–1935 is divided by GDP₁₉₃₅ to calculate the fiscal variable ratios of 1935. An alternative method would be to calculate, as Mata (1993) did, an average GDP between t and t+1. However, this alternative method would entail having a break in 1936.

We also used official data for the whole period, without making any adjustments to it (except correcting for the recent one-off measures). As pointed out by Bohn (1991), the official data is most informative about government behavior if policy-makers are primarily influenced by the official data. The main source of data is the publication from the Portuguese Institute of National Statistics (INE) coordinated by Valério (2001). For the period before 1913 the original data was collected by Mata (1993). Since this data is less rounded we have used the original source. Detailed series for the period 1913–1947 are available in Valério (1994). All variables were converted into euros using the irrevocable conversion rate of 1EUR = 200.482 PTE. The detailed source for each variable is given below.

As discussed in the text, our data refers to the central government accounts, expressed on a public accounting basis and not on a national accounting basis. This is because the main source of the historical data is the “Conta Geral do Estado”, which is a yearly publication from the Ministry of Finance containing the final information on budget execution, in a public accounting perspective. Data on a national accounting basis is only available from 1947 onwards. A coherent dataset, compiled by the Bank of Portugal for the period 1947–1995, is published in Pinheiro (1999). The main drawback of our dataset is the non-inclusion of the other sub-sectors of the public administrations. It is not possible, however, to find historical data for the whole sample period. In contrast, as public accounting mostly follows a cash basis registering, it is theoretically more compatible with the debt series.

Detailed sources

GDP

1861–1994: Valério (2001). The source for this variable for the period 1947–1994 corresponds to Pinheiro (1999), i.e. the author has linked the historical series with the Bank of Portugal’s estimates for the 1947–1995 period. This series was adjusted to the level of the next period, which is in the ESA95 standard.

1995–2003: INE estimates, as they appear in Ministério das Finanças (2003).

Primary expenditure

The primary expenditure corresponds to the difference between total effective expenditure, total effective public expenditure and interest payments on the debt. The source of actual data on total effective expenditure was:

1852–1913: Mata (1993); 1914–1998: Valério (2001)

1995–2003: collected by us using the same criteria, from the Conta Geral do Estado (yearly publication).

Public revenues

Corresponds to total effective public revenues of the Portuguese State (Estado) adjusted for the effects of the extraordinary revenues obtained in 2000, 2002 and 2003. Such excluded revenues amounted to 399 million EUR in 2000 (UMTS); 1830 millions in 2002 (CREL revenues, sale of the fixed telecommunication network to Portugal Telecom and an extraordinary regularization of taxpayers’ debts); and 1962 millions EUR in 2003 (mainly sale of government credits to Citigroup, and integration of the pension fund of the Portuguese post office, CTT, into the public servants’ pension system, CGA). The effective revenues exclude the revenues from loans.

1852–1913: Mata (1993); 1914–1998: Valério (2001)

1999–2003: collected by us using the same criteria, from the Conta Geral do Estado (yearly publication).

Budget balance

Corresponds to the difference between total effective public revenues and total public expenditure.

Public debt

1861–1913: Mata (1993); 1914–1979: Valério (2001)

1980–2003: Instituto Gestão do Crédito Público (IGCP), Direct State Debt

Interest payments on the debt

1861–1913: Mata (1993), Table 11. The author distinguishes total public debt servicing costs, which comprises amortization of the debt, interest payments and administrative costs. There are observations on the interest missing for the years 1860, 1866, and 1898–1907. We assumed that the interest payments in 1860 were an average of its values in 1859 and 1861. Due to a lack of data on administrative costs, the interest data for 1866 is calculated by difference, assuming that the administrative costs remain equal to their average weight in total during the last three years before 1866 (0.35%). From 1898 to 1913 Mata (1993) is again unable to disaggregate between administrative costs and interest payments. We again found the interest payments by computing the difference between the total servicing costs, excluding the amortization of the debt, and the estimated administrative costs, which were obtained assuming that their weight in total is equal to the average for the 3 years before and the 3 years after the lack of data (0.3%).

1914–1947: Valério (1994), table 22

1948–2003: collected by us, from the Conta Geral do Estado (yearly publication).