Ecuador: Selected Issues and Statistical Annex

The crisis in the banking sector was one of the major contributing factors that led Ecuador to abandon its own currency and introduce the U.S. dollar as legal tender. However, to illustrate the weak growth performance of the country, it is necessary to examine the structural weaknesses in the labor market, the tax system, and the trade system. These weaknesses resulted in the increase in poverty and inequality. This paper provides a brief summary of recent economic developments and statistical data on economic indices of Ecuador.

Abstract

The crisis in the banking sector was one of the major contributing factors that led Ecuador to abandon its own currency and introduce the U.S. dollar as legal tender. However, to illustrate the weak growth performance of the country, it is necessary to examine the structural weaknesses in the labor market, the tax system, and the trade system. These weaknesses resulted in the increase in poverty and inequality. This paper provides a brief summary of recent economic developments and statistical data on economic indices of Ecuador.

II. Sources of Economic Growth, 1970-99 4

18. This chapter provides an overview of Ecuador’s economic growth over the past three decades. The objective is to highlight some of the key features of long-term growth, and thus to set the stage for the subsequent chapters. The discussion is divided into three sections: the first looks at Ecuador’s growth in a regional perspective, the second uses a production function approach to calculate the contribution to growth from capital, labor, and total factor productivity, and the last looks more closely at growth in the 1990s.

A. Growth Performance in a Regional Perspective

19. A simple comparison of economic growth per capita in Ecuador with seven other countries in the region is provided in Figure II. 1 and Table II. 1.5 The difference between Ecuador and the other countries in growth performance is striking:

Figure II. 1.
Figure II. 1.

Ecuador: Per Capita GDP Growth of Selected Latin American Countries

(Index numbers, 1970=100)

Citation: IMF Staff Country Reports 2000, 125; 10.5089/9781451811773.002.A002

Source: World Economic Outlook
Table II. 1.

Ecuador: Comparison of Growth Indicators for Eight Latin American Countries, 1970-99

article image
Source: World Economic Outlook
Table II. 2a.

Ecuador: Production for the Oil Sector, 1965-93

article image
Source: Fund staff estimates.
Table II. 2b.

Ecuador: Production for the Non-Oil Sector, 1965-93

article image
Source: Fund staff estimates.
  • Ecuador had strong growth from 1970 until 1981; during this period GDP per capita increased by 70 percent, more than any other country in the sample. Since then, however, GDP per capita has effectively been stagnant—first declining in the early to mid-1980s, and then growing slowly in the 1990s until the present crisis—resulting in a level in 1997-98 that was the same as in 1981.

  • In contrast, Argentina, Bolivia, Chile, and Peru went through the 1970s and early 1980s without any gains in GDP per capita. Since then, however, these four countries have recorded sharp recoveries; a doubling of GDP per capita in Chile between 1985 and 1999, and overall increases of 25-35 percent since around 1990 in the others. Colombia had steady improvement in GDP per capita; about 70 percent from 1970 to 1999. Venezuela recorded a decline of about 15 percent during the same period.

20. In 1998, Ecuador had a GDP per capita, measured in PPP terms of US$3,003; only marginally higher than Bolivia, and about one fourth of that in Argentina.

21. However, the differences between Ecuador and the other countries in the sample are far less pronounced in terms of conventional factors of growth, e.g., accumulation of physical and human capital. The average rates of investment and domestic savings relative to GDP in Ecuador have been at a broadly similar level to those of the other countries for the whole period. Moreover, the improvements in a few selected human development indicators have been similar to the other countries (see Table II. 1).

22. The reasons behind the turnaround in the mid-1980s for Argentina, Bolivia, Chile, and Peru have been researched extensively and documented elsewhere: a change from unsustainable policies and severe macroeconomic instability, culminating with the debt crisis in the early 1980s, to the subsequent introduction of more stable macroeconomic policies combined with structural reforms. Ecuador’s different experience can, to some extent, be ascribed to exogenous shocks. The strong growth in the 1970s was driven in large part by discoveries of new petroleum fields in the eastern region of the country and the sharp increase in the international price of oil in 1972-73. Similarly, the decline in oil prices in the early 1980s and the devastating earthquake in 1987 clearly dampened growth. However, there is also ample evidence, discussed in later chapters, that Ecuador’s failure to pursue sound financial policies in combination with the lack structural reforms has been another important factor behind the disappointing performance, especially in the 1990s.

B. Sources of Growth in Ecuador

23. This section provides estimates of the contributions from capital, labor, and total factor productivity to potential output growth for Ecuador during 1975-93.6 First, production functions for two separate sectors (“oil” and “non-oil”) are estimated econometrically, using an error correction model.7 These production functions provide the basis for a growth accounting framework that allows us to calculate total factor productivity (TFP) and the contribution to growth from capital, labor, and TFP (see Annex).

24. The estimation results for the two production functions—oil and non-oil sectors—are presented in Tables II.2a and II.2b.8 In both sectors the dependent variable is the log of the ratio of sectoral GDP to sectoral labor input. Three general results apply to both sectors:

  • The hypothesis of constant returns to scale could not be rejected

  • Human capital measures were not statistically significant as inputs, and are, therefore, not reported;

  • A time trend intended to capture the possibility of disembodied technical progress was not significant. Therefore, technical progress seems to have been embodied in the physical capital.

25. For the oil sector, the regressors are a constant, and four other variables which interacts with a step dummy that equals one for 1974 onwards and zero otherwise: the constant, the change in the log of capital to labor, the log of the ratio of GDP to labor lagged, and the log of the ratio of capital to labor lagged. The step dummy for 1974 was statistically significant, indicating that the existence of a structural break after the exogenous shocks of years 1972 and 1973, and thus that a change in the key parameters of the sector could not be rejected. Similarly, the three impulse dummies for the years 1972, 1973, and 1987 were statistically significant, accounting for shocks to the sector.

26. For the non-oil sector, the regressors are a constant, the change in the ratio of capital to labor, the log of the ratio of GDP to labor lagged, the log of the ratio of capital to labor lagged, and the lead of the change in the log of the ratio of capital to labor. A step dummy variable from 1972 onwards was included, and was significant, providing some evidence of a positive spillover effect from the oil sector to the non-oil sector.

27. The calculation of the contributions to growth is summarized in Table 3. The Ecuadoran economy grew at an average annual rate of 3.6 percent in real terms in 1975-93 (3.4 percent in the non-oil sector, and 4.8 percent in the oil sector). Four points are noteworthy from Table II. 3:

Table II. 3.

Ecuador: Sources of Growth, 1965-93

article image
Sources: Central Bank of Ecuador; National Statistics Institute; and Fund staff estimates
  • Since the mid-1970s, i.e., since immediately after the big oil windfall, there has been a steady decline in Ecuador’s potential GDP growth rate; from 6.4 percent annually in 1975-80 to 1.8 percent in the early 1990s.

  • The capital contribution to growth was quite large, i.e., 6-7 percent in both sectors, during the period 1975-80, but has since declined sharply.

  • The contribution of labor to growth has been fairly steady in both sectors throughout the observation period.

  • The TFP contribution has, for the whole economy, been negative throughout the observation period.9 For the non-oil sector, the TFP contribution was positive only during 1965-75, i.e., in the period before the oil windfall. For the oil sector, the TFP contribution was positive only during the early nineties, when the sector grew rapidly despite no new investments.

28. This simple growth accounting framework obviously cannot establish what “caused” economic growth. What it does establish, however, is that TFP has been significantly less of a factor behind growth in Ecuador than in other countries in the region; typically periods of high growth are associated with periods of high TFP contribution to growth (Table II. 4). Total factor productivity is a broad measure of the flexibility of an economy and thus its ability to allocate a growing stock of productive resources to its most efficient uses. This suggests that there are important structural rigidities in the Ecuador economy that inhibit a more vigorous growth response.

Table II. 4.

Ecuador: Sources of Growth—International Comparison

article image
Sources: Elias (1992); Abreu and Verner (1997); and Fund staff calculations.

C. Growth in the 1990s

29. Data limitations do not permit the above detailed analysis to be carried beyond 1993. However, a closer examination of national accounts data for the period 1990-99 reveals some interesting observations:

  • Trend growth, illustrated in Figure 2 by applying a Hodrick-Prescott (HP) filter to real GDP growth, was negative throughout the decade. This conclusion is fairly robust to the choice of end-point (i.e., by eliminating a possible bias from the strong negative growth in 1999), to using annual or quarterly GDP figures, and to using moving averages rather than a HP filter to calculate underlying trend. This replicates the finding above, and suggests that potential, or trend, growth has been on a declining path since around 1975, and ending the decade at about 1.6 percent annually. With an annual population growth averaging 2.5 percent, a rather significant reversal of trend growth would be necessary to improve average living standards.

  • An important reason for the decline in growth in this decade has been the declining contribution to growth from the petroleum sector (Table II.5). In 1998-99 this decline is obviously related to the sharp drop in oil prices in 1998 following the Asian crisis. However, the declining growth contribution started earlier in the decade, and is explained mainly by the fiscal constraint: maintenance expenditures on existing oil installations have been negligible for several years, resulting in declining production volumes.

  • The contribution to non-oil growth in the 1990s has come predominantly from non-tradeables sectors, i.e., construction, and services (except financial services), while tradeable sectors, especially manufacturing and to some extent agriculture, have contributed a sharply declining share (Table II.5). A similar picture emerges when looking at the contribution to growth by expenditure components: the contribution from domestic demand increased sharply in 1997-99. To some extent, this illustrates the impact of the El Niño weather phenomenon that devastated much of coastal agriculture in 1997. However, the growing contribution from the service sectors also suggests a bias in underlying economic incentives that have attracted more resources toward the non-tradeable sector, and thus to a more “inward-looking” economy.

Table II. 5.

Ecuador: Contributions to Growth, 1991-99

article image
Source: Fund staff calculations.
Figure II. 2.
Figure II. 2.
Figure II. 2.

Ecuador: Actual and Trend GDP Growth

Citation: IMF Staff Country Reports 2000, 125; 10.5089/9781451811773.002.A002

Sources: Ecuadoran authorities; and Fund staff estimates.

• Annex: A Production Function Approach

A Cobb-Douglas specification of the production function (in logs) can be written as:

(1)Log Yt = a+bt+α Log Kt+β Log Lt+(1αβ) Log Ht

If neither the trend nor the human capital term is significant the production function reduces to the intensive form:

(2)Log yt=a + α Log kt

where y=Y/L and k=K/L.

The production function equation was estimated using cointegration techniques that are particularly appropriate for uncovering the long-run relationships between inputs and outputs.10 Once α is estimated, a total factor productivity series (in logs) can be calculated as:

(3)Log At=Log Ytα Log Kt(1α) Log Lt

Neither the capital stock nor the labor measure (which corresponds to the economically active population concept) is adjusted by the degree of utilization or unemployment. This does not affect the estimates of the capital share, and thus of the long-run contribution of capital and labor. However, it may affect the estimates over shorter horizons. One way to address this problem is to use “smoothed,” or trend, estimates of TFP, using the Hodrick-Prescott filter. By replacing this smoothed TFP, together with the estimate of the capital share in equation (2), an estimate of potential output is calculated.

4

Prepared by Marcio Ronci, Marcelo Sanchez, and Erik Offerdal.

5

The other countries in the sample are: Argentina, Bolivia, Brazil, Chile, Colombia, Peru, and Venezuela.

6

The analysis could not be carried beyond 1993 because of lack of reliable data on capital stocks after this date.

7

The oil sector is defined to include all of division 2 in the ISIC classification; it thus corresponds to the mining sector. In Ecuador, division 2’s dominant components are crude oil and gas; it also includes refining of petroleum and other mining.

8

The test for cointegration is based on the OLS coefficient of the lagged dependent variable in an autoregressive distributed lagged model (eventually) augmented with leads of the regressors. For an application of this modeling strategy, see Belke and Golke (1996).

9

Series of TFP growth in both of the sectors were estimated by subtracting a weighted sum of factor inputs from GDP growth, using as weights the factor share estimated in the regressions in Tables II.2a and II.2b. The TFP series were then smoothed using a Hodrick-Prescott filter.

10

Since it is possible to find the presence of structural breaks during the sample period, we have tested for stability of the estimating equations as well as tried several step and impulse dummies.

Ecuador: Selected Issues and Statistical Annex
Author: International Monetary Fund