Journal Issue

Pakistan: Selected Issues and Statistical Appendix

International Monetary Fund
Published Date:
November 2005
  • ShareShare
Show Summary Details

I. Overview

1. Pakistan’s economic performance continues to be strong. Economic growth exceeded 8 percent in 2004/05—the third year of high growth. Fiscal adjustment, supported by official and private inflows and debt relief, has led to a substantial improvement in public and external debt indicators. International reserves have recovered to close to $10 billion. Financial sector reforms have resulted in a healthy banking system. With these achievements, vulnerabilities have been greatly reduced and Pakistan’s prospects look favorable. But will it be just smooth sailing ahead, or are there still challenges to be tackled? This Selected Issues volume tries to shed some light on this.

2. Raising investment will be key to maintaining high rates of growth in years ahead. Chapter II compares the current growth acceleration in Pakistan with two earlier periods of strong growth. Each of these were preceded by—or coincided with—a substantial increase in the investment ratio and occurred during periods of relatively low inflation. The current growth acceleration has not seen much of an increase in the investment ratio, at least not according to national accounts data. The increase in total factor productivity witnessed in the last few years may have reflected in part a reduction in existing excess capacity, following the slump of the late 1990s, as well as favorable weather conditions that boosted agricultural production. With many sectors of the economy approaching full capacity, it is now necessary to move the production frontier outward through a rise in investment.

3. Pakistan’s debt burden has declined sharply in the last four years, but remains moderately high. Chapter III finds that Pakistan’s debt situation today stands in sharp contrast to the late 1990s, when Pakistan faced difficulties in meeting some of its debt service obligations. Pakistan’s public debt declined by almost 30 percent of GDP since its peak in 2000/01. But with debt still at 60 percent of GDP, risks remain, particularly as the debt-to-revenue and debt-to-exports ratios are still high. A continuation of prudent fiscal policies, as anchored by the recently adopted financial responsibility law, is needed to ensure that debt ratios continue on their downward trajectory.

4. The banking sector has continued to strengthen. The 2004 Financial System Stability Assessment (FSSA; IMF Country Report No. 04/215) found that banks’ financial health had greatly improved following several years of restructuring and privatization. Chapter IV provides an update on financial sector developments since the FSSA. It finds that most financial soundness indicators continue to show an improving trend. However, lending activity appears to have approached “boom” thresholds, underscoring the need for continued supervisory vigilance. The stock market, after showing stellar gains in the last few years, underwent a major correction in early 2005. This fortunately did not have a significant impact on banks’ health or on the economy in general, but it does point to the need for further strengthening of securities markets regulation and oversight.

5. The corporate sector has also witnessed a sharp recovery in recent years. Chapter V examines the vulnerability of Pakistan’s corporate sector, looking at the sector’s governance structure and a score of balance sheet indicators. It finds that, as with the banks, the sector’s financial health has improved considerably over the past few years. The corporate governance structure has improved as well, although strong family ownership and control could become a constraint to faster expansion. Pakistan’s corporate sector relies mostly on internal and bank financing, and much less on shares or bonds, and its external exposure is very limited. Bank financing has been growing rapidly and warrants careful monitoring. Valuation indicators, meanwhile, have remained relatively low, suggesting that there is still a perception of high risks attached to doing business in Pakistan. These would not only include political and security risk, but also remaining governance problems. While progress has been made in this area, reforms to improve the business environment in Pakistan will therefore need to continue.

II. Is Pakistan’s Growth Acceleration Sustainable?1

A. Introduction

6. In the past few years, Pakistan’s economy has made a sharp recovery, moving from economic crisis to strong growth. In the second half of the 1990s, growth rates had fallen to an average of 3 percent per year, barely exceeding population growth. The government that came to power in 1999 put macroeconomic stabilization and key structural reforms at the top of its agenda. With sound fiscal and monetary policies, and aided by strong international support, the economy has witnessed a dramatic turnaround, with growth accelerating to over 8 percent in 2004/05. The question that arises is whether these high growth rates can be sustained in the coming years, as the government aims to achieve.

7. Raising the rate of growth in a sustained manner has been much discussed in economic theory and policy. There is a vast body of literature that attempts to answer the question of how to promote growth (see, for example, Barro (1991), Barro and Sala-i-Martin (1994), Bosworth and Collins (2003), and Easterly (2001)). But despite the voluminous literature, empirical analyses of country experiences do not offer reliable and unambiguous results to answer this question.

8. The good news, according to recent literature, is that growth accelerations are a fairly frequent phenomenon. Hausmann, Pritchett, and Rodrik (2004) find that a country has a one-in-four chance to experience a growth acceleration sometime during a decade, with an acceleration defined as real per capita growth of 2 percent or more lasting for at least eight years. They also find that growth accelerations tend to be correlated with increases in investment and trade, with real exchange rate depreciations, and with political regime changes. The bad news is that not all accelerations are sustained. External shocks, for example, tend to produce growth accelerations that fizzle out, but economic reform is a significant predictor of accelerations that are sustained.

9. Growth accelerations tend to be highly unpredictable. The vast majority of growth takeoffs are unrelated to the standard determinants found by Hausmann et al. (2004), such as increases in investment and trade, and growth takeoffs typically fail to materialize when these conditions are indeed favorable. Similarly, Rodrik (2003) argues that igniting economic growth and sustaining it are somewhat different enterprises. Again some good news, as he finds that it often takes only small reform steps to stimulate growth. But it requires continued institutional reforms to sustain growth, by improving resilience to shocks and maintaining productive dynamism. Rodrik emphasizes that there are a few first order economic principles that need to be adhered to—protection of property rights, market-based competition, appropriate incentives, and sound money—to maintain strong growth. These principles can translate into very different policy packages, however, for individual countries. Reformers have substantial room for creatively packaging these principles into institutional designs that are sensitive to local opportunities and constraints.

10. Economic growth in Pakistan has averaged a very respectable 5½ percent per year since 1960. Real per capita GDP has been growing at an average of about 2¾ percent per year, meaning that real incomes have more than tripled during this period. Since 1960, Pakistan has experienced two earlier sustained growth accelerations, with per capita real growth rates consistently exceeding 2 percent per year, one that started in 1961 and one in 1977, and lasting 10 and 12 years, respectively (Figure II-1).

Figure II.1.Pakistan: Economic Growth, 1960–2004

11. Estimating trend growth using a Hodrick-Prescott filter suggests that the Pakistani economy has entered another period of strong growth (Figure II-2). It also suggests that growth is currently above trend level. But this does not tell us, of course, how long this upswing will last, or, in other words, when the next turning point might be (although econometric techniques exist to help forecast turning points). This chapter tries to discern some of the factors underlying the earlier growth spurts and ascertain if they can provide some insights as to whether the current growth acceleration might prove sustainable.

Figure II.2.Pakistan: Actual and Trend Economic Growth, 1960–2004

B. Growth Accounting

12. To analyze Pakistan’s past growth experience, as a first step a standard growth accounting approach is used. Using a simple two-factor production function that relates output (Y) to the quantities of capital (K) and labor (L):

A is total factor productivity (TFP) and α; is the elasticity of output to changes in capital. This can be re-written in terms of changes to decompose the rate of output growth (y) into the contribution of growth in the inputs, k and l, plus the change in TFP, a (which is essentially a residual capturing a multitude of factors, as noted also below in Section C):

In the absence of reliable time series for the capital stock, the gross investment rate can be used as a proxy for the change in the capital stock. The change in the capital stock is equal to:

where δ is a measure of the rate of depreciation. Dividing through Kt-1 and assuming a steady-state constant value γ for the inverse of the capital-output ratio allows the rate of change of capital (k) to be measured by the investment rate (i = It/Yt):

Replacement of k with its steady state approximation yields the formulation used in many growth studies:

In line with many other studies on growth, the output elasticity of capital is assumed to be 0.35 and the rate of depreciation 5 percent (see, for example, Senhadji (2000)). The steady-state capital ratio is calculated using estimates for Pakistan’s 1990 capital stock from Crego (1998) and that year’s GDP, resulting in a value for γ of 0.7. It should be emphasized that because of these assumptions, and data limitations in general, the results presented here are more indicative than exact measures.

13. In both the two earlier periods of sustained strong growth, growth resulted from an increase in capital inputs, as well as an increase in TFP (Table II.1). Indeed, the two previous sustained growth accelerations were preceded by—or coincided with—a significant increase in the investment ratio. In the early 1960s, the investment ratio2 rose from just over 12 percent of GDP in 1960 to 22½ percent in 1964 (Figure II.3). By 1971, when this ten-year period of strong growth ended, the investment ratio had declined again to about 14 percent of GDP. Similarly, the investment ratio rose sharply to 19 percent of GDP in the two years preceding the 1977 growth acceleration. As periods of sustained growth have been defined here as lasting as long as real per capita growth remains at least 2 percent each year, this second period ended in 1992. However, following a dip in 1993, growth was fairly strong again in 1994–96. Starting in 1993, the investment ratio started to decline, falling back again to 14 percent by 1998.

Figure II.3.Pakistan: Economic Growth and Investment, 1960–2004

Table II.1.Pakistan: Growth Accounting, 1960–2004
Real GDP growth5.386.324.946.254.464.486.566.366.533.75

14. The recent growth acceleration was not preceded or accompanied by a similar increase in the investment ratio. The contribution of capital to growth is actually below the historical average and well below that of the two previous growth spurts. According to the latest available estimates, the investment ratio has remained fairly stable at about 15½ percent of GDP since 1999. There are indications, however, that investment may be underreported in the national accounts—as suggested by the sharp increases in commercial credit and imports of machinery—but it is unclear to what extent and whether this is different from earlier years (Table II.2).

Table II.2.Pakistan: Investment Indicators, 2000/01–2004/05
(In percent of GDP)
Gross fixed capital formation15.815.515.315.615.3
Change in inventories1.
Gross capital formation17.216.816.917.316.8
Import of machinery3.64.45.3
Foreign direct investment0.
Commercial credit14.913.914.817.119.0
Source: Pakistani authorities.
Source: Pakistani authorities.

15. The recent growth acceleration has come largely from an increase in TFP. The contribution of TFP to growth in the last few years is similar, or even somewhat higher, than in the earlier growth periods. To some extent, this may reflect the growing contribution of the services sector to growth, which is likely to require less investment compared to manufacturing. It also may reflect that following the slump of the late 1990s there was considerable excess capacity in the economy and therefore less of a need for new investments to generate growth. The increase in capacity utilization translates into higher productivity per unit of capital and is reflected in the higher-than-average contribution of TFP to growth. But with many sectors approaching full capacity, sustaining growth in the coming years would now appear to require an increase in the investment ratio, as without new investments it might be difficult to continue to improve productivity at the same pace as in these last few years.

16. With agriculture accounting for about one quarter of GDP, it is worth analyzing the impact of variations in rainfall on growth. Agricultural production in Pakistan is heavily dependent on rainfall because of the limited infrastructure to smooth the supply of water. Including the change in rainfall, r, in production function (5) gives:

Rainfall data were obtained from the Federal Bureau of Statistics of Pakistan, by taking the average of annual rainfall in Lahore, Peshawar, and Rawalpindi/Islamabad.3 Assuming, admittedly rather arbitrary, a small weight for rainfall (0.05) does not change the growth accounting picture much for the entire period 1960–2004 (Table II.3), nor for the earlier periods of sustained growth. But it does have a major impact on the picture for the recent growth acceleration. The results suggest that improved rainfall has been a significant contributor to the recent growth acceleration. This is consistent with the increase witnessed over the last four years in the contribution of agricultural production—and particularly of that of major crops such as cotton and wheat—to the overall growth rate, which improved by almost 2½ percentage points, compared to a similar increase in the contribution of the services sector and an increase in the contribution of the industrial sector of 1½ percentage points (Table II.4).

Table II.3.Pakistan: Growth Accounting, 1960–2004
Real GDP growth5.386.324.946.254.464.486.566.366.533.75
Table II.4.Pakistan: Contributions to GDP Growth
Of which:major crops-
Of which:manufacturing1.42.20.8
Of which:commerce0.82.21.4
GDP growth at factor costs1.88.46.5
Source: Pakistani authorities.
Source: Pakistani authorities.

C. Growth Determinants: Some Stylized Facts

17. This section goes beyond the growth accounting framework and tries to identify broader factors that may determine growth in Pakistan. Growth accounting is a fairly mechanical approach, subject to various criticisms. A first concern is that the results are sensitive to underlying assumptions about the nature of the production function and to the indicators chosen to measure changes in outputs and inputs. Also, TFP is measured only as a residual. It provides a measure of economic efficiency, that is, the quantity of output that can be produced with a given quantity of inputs or, in other words, an outward shift of the production function (Easterly and Levine, 1997). Such gains in efficiency can reflect a myriad of factors that influence growth, but which the measured increases in factor inputs do not account for. In addition to technical innovation, TFP also reflects, for example, political (in-)stability, economic policies, or institutional changes, which affect the efficiency of an economy in much of the same way as technology does. But most of all, an accounting decomposition does not determine the fundamental causes of growth (Bosworth and Collins, 2003). It merely provides a framework for identifying the proximate sources of growth. To get a better understanding about the determinants of growth, a simple regression analysis was conducted, with the results presented in Table II.5. The dependent variable is real GDP growth. The explanatory variables include lagged growth, the rate of investment, changes in annual rainfall, the rate of inflation, world growth, and real exchange rate changes. The sample period is 1960–2004.

Table II.5.Regression Results
Dependent variable: real GDP growth, 1960-2004
Real GDP growth, lagged one year (percent)0.520.51
Investment ratio, lagged one year (percent of GDP)0.170.16
Average rainfall current and previous year (percent change)0.0180.018
Inflation, lagged one year (percent)-0.064-0.069
Real exchange rate change, lagged one year (US$, percent)-0.028
Partner country growth, lagged one year (percent)0.13
Adjusted R-squared0.370.33
Notes: Estimation is by OLS. T-statistics in parentheses.
Notes: Estimation is by OLS. T-statistics in parentheses.

18. The overall explanatory power of the regression results is limited, highlighting the difficulty of pinpointing specific determinants of growth (Figure II.4). It also seems to confirm the broad consensus emerging from the literature that little is still known about what determines growth. Only a little over one-third of the variation in growth is explained by the determinants in Table II.5. Of course, this also reflects the very limited availability of consistent time series for the entire sample period, for example indicators describing the quality of institutions.

Figure II.4.Pakistan: Actual and Fitted Growth Rates

(equation 1), 1963–2004

19. The results do confirm the importance of investment and rainfall as determinants of growth in Pakistan. The coefficients for both investment and rainfall are statistically highly significant and have the expected sign. It should be noted that growth and investment may depend on the same set of fundamentals. The lagged value of the investment ratio was used in an attempt to overcome this problem, although it is recognized that this does not eliminate this problem altogether. Rainfall is a major determinant of agricultural output, and thus, to a lesser extent, of overall output. Changes in the average rainfall of the current and previous year explain about 70 percent of the variation in agricultural value added (Figure II.5). The agricultural sector in turn accounts for about one quarter of overall GDP and also has significant indirect effects, for example on the textile industry. Rainfall is, of course, a factor the authorities have no control over, but they can work to improve water management. Indeed, the government is now planning major investments in this area in order to reduce the impact of fluctuations in rainfall.

Figure II.5.Pakistan: Agricultural Growth and Rainfall, 1960–2004

20. Macroeconomic stability also appears to be a (pre-)condition for growth. The coefficient for inflation is significant and has the expected sign. Periods of sustained growth appear to have been preceded by a reduction in inflation from relatively high levels (Figure II.6). Prior to the 1977 growth spurt, inflation was reduced to 7 percent, down from a peak of almost 27 percent in 1974. Similarly, albeit less dramatically, inflation was reduced to 4 percent in 1999, following a peak of about 12 percent in 1994–95. Inflation averaged 3½ percent during the growth spurt of the 1960s, 7½ percent during the one starting in 1997, and almost 5 percent during 2003–04. By comparison, inflation averaged 15 percent during 1971–76 and 10½ percent during 1991–98.

Figure II.6.Pakistan: Economic Growth and Inflation, 1960–2004

21. A depreciation of the real exchange rate tends to result in higher growth. While world demand was not found to have a statistically significant effect on growth, real exchange rate depreciations tend to be correlated with stronger growth. Pakistan is not an economy that is highly integrated into the world, based on the relatively low ratio of trade to GDP, but this underscores the export industry’s importance for the country’s overall economic performance. It also suggest that the State Bank of Pakistan is correct in monitoring real exchange rate developments closely and aiming to avoid a real appreciation.

D. Concluding Remarks

22. From the above findings, it is not clear that the current growth acceleration will be sustained. While the results of this paper confirm that growth remains highly unpredictable, it does suggest that an increase in investment—including in water management—would be needed for growth to remain strong. The increase in TFP witnessed in the last few years may have reflected in part a reduction in excess capacity, as well favorable weather conditions that boosted agricultural production. With many sectors of the economy approaching full capacity, it now becomes necessary to move the production frontier outward through new investments. Meanwhile, inflation has picked up and, if left unchecked, could undermine confidence and thus investment and growth. With the continuation, however, of the authorities’ sound economic policies and structural reforms of the last few years, the business environment will continue to improve and inflation should return to lower levels. Several institutional improvements have been implemented that should help ensure that policies remain on the right track: the central bank has been strengthened, the financial system’s health has improved, and a fiscal responsibility law was adopted to ensure fiscal sustainability. Thus, investment can be expected to increase in the coming years, helping to maintain strong economic growth.


Prepared by Ron van Rooden.

Gross fixed capital formation, including both private and government investment.

The average of these three cities gives an average annual rainfall for the 1960–2004 period of about 750mm, which is similar to average annual rainfall figures quoted in general descriptions of the country.

Other Resources Citing This Publication