Journal Issue
Share
Article

Saudi Arabia: Selected Issues

Author(s):
International Monetary Fund. Middle East and Central Asia Dept.
Published Date:
July 2013
Share
  • ShareShare
Show Summary Details

Productivity Growth and Potential Ouput in Saudi Arabia1

The Central Department of Statistics (CDSI) has recently released revised national accounts data that significantly changed the path of real GDP growth over the past decade. Average overall GDP and non-oil GDP growth from 2005–11 are now estimated at 6.4 percent and 8.4 percent per annum in real terms, respectively, compared to 3.9 percent and 5.3 percent previously. The improved coverage of the non-oil sector is the main reason for the revisions. In light of the substantial upward revision to the GDP data, this paper revisits the issue of productivity growth in Saudi Arabia, and reassesses estimates of medium term potential growth, especially for the non-oil sector. The paper has three main conclusions: (i) labor productivity and TFP growth in the non-oil economy over 2005–11 have been stronger than previously thought, at around 2.5 percent and 2.2 percent per annum; (ii) nevertheless, capital accumulation remains the principal driver of growth in the non-oil sector; and (iii) both time-series filters and a growth accounting methodology suggest that potential growth in the non-oil sector is around 5-6 percent.

A. Dynamics of Output and Factor Inputs

1. Non-oil sector growth accelerated significantly after 2000, averaging over 6.3 percent annually, compared to less than 2.7 percent in the 1990s.2 The private sector was the key driver behind the stronger non-oil sector growth, with an annual growth rate close to 7 percent after 2000, while growth in the public non-oil sector was about 4.7 percent on average during the same period. In contrast, oil sector growth was the most volatile series among the three, and averaged only 2.3 percent (Figure 1). As a result, the non-oil sector is playing an increasingly important role in driving economic growth in Saudi Arabia in recent years (Figure 2).

Figure 1.Real Output Growth, 1990–2012

Source: CDSI.

Figure 2.Sector Shares in Real GDP, 1990–2012

Source: CDSI.

2. The acceleration in growth in the private sector, however, could be partly due to the way the GDP data has been revised. The revised data is only available from 2004 onward. Prior to this, staff has used the growth rates of the old GDP series to splice together a continuous GDP series back to 1990. However, since the revision shows a substantial rise in the non-oil sector growth rate (Figure 3), this could suggest under-recording of non-oil GDP growth in the earlier years.

Figure 3.Non-oil Growth Revision, 2000–11

Source: CDSI.

3. Factor inputs have been generally increasing to support growth, although at different rates. Not surprisingly, they had the most robust growth in the private sector (Figures 4 and 5). The increase in labor input in the private sector since 2000, growing 5.4 percent per annum on average, was mainly driven by the increasing number of expatriate workers. While public sector employment was stable throughout the 1990s, it started to pick up in 2000. Similarly, public investment was declining in real terms prior to 2000, but has grown quickly since. Indeed, real public investment increased more than six fold between 2000 and 2011. In the oil sector, investment had been relatively low, rose substantially in 2006–07, and has remained flat thereafter. Employment growth in the oil sector has been steady, although it is low relative to other sectors. It should be noted, however, that the official data may underestimate employment because of the presence of illegal immigrant workers in Saudi Arabia. This could affect the results of the analysis both in terms of the level and growth dynamics.

Figure 4.Employment, 1990–2012

(Millions)

Sources: CDSI; SAMA; and IMF staff estimates.

Figure 5.Real Investment, 1990–2011

(Billions of 1999 SARs)

Sources: CDSI; SAMA and IMF staff estimates.

B. Investigating Productivity Growth in the Saudi Arabian Economy

Labor productivity, measured by unit labor real output, has picked-up since 2005 in the non-oil sector, but has been broadly flat in the economy as a whole. Labor productivity for the whole economy has remained more or less flat since 1990, while it dropped significantly in the oil sector. In contrast, labor productivity in the non-oil sector was broadly unchanged during the 1990s, but started to grow thereafter, driven by improvements in private sector productivity (Figure 6). While the analysis in this paper does not look at the factors that could be driving productivity growth in Saudi Arabia, the WTO accession in 2005, the stock market liberalization during the 2000s, and increased government infrastructure spending since 2004 could all have played a role, although statistical issues with the data revisions and possible under-recording of employment data could also be at play. In the analysis, estimates for the public sector are provided for completeness, although given the way value-added in the public sector is estimated in the national accounts, interpretation is difficult. For example, unit labor output in the public sector rose substantially in 1990s, but this reflected real wage increases rather than an improvement in underlying productivity.

Figure 6.Unit Labor Output, 1990–2012

(Index, 1990 = 100)

Sources: CDSI; SAMA; and IMF staff estimates.

4. Estimates of TFP show little improvement for the whole economy, with a deteriorating trend in the oil sector and a significant improvement in the non oil sector in recent years mainly due to the private sector. A growth accounting approach that takes into account factor accumulation and provides a quantitative assessment of TFP growth is used below.3 Starting with a standard Cobb-Douglas production function:

where α is the cost share of capital. Initially, α=23 is assumed across all sectors to reflect the fact that capital in emerging and developing economies tends to be scarce and thus earns a relatively high rate of return. The quantitative assessment of TFP could be sensitive to the value of this parameter. Therefore, the parameter value is then estimated using sector level data in the second phase, and these sector specific estimates are applied to the growth accounting framework to assess TFP growth.

TFP estimates assuming α=23

5. For the economy as a whole, TFP growth averaged less than 0.6 percent per year over the period. TFP in the oil sector has been declining at a rate of 1.7 percent a year, and by even more in recent years (Figure 7). The estimates show a positive average TFP growth rate of 1.3 percent in the non-oil sector. Private sector TFP grew quickly in recent years, averaging 3.4 percent since 2005, although it averaged close to zero (-0.3 percent) over the whole period. For the public sector, the average growth rate of TFP was 2.4 percent. It mostly comes from the period of near zero employment growth and reduced capital stock due to the subdued investment in the public sector. Estimates of TFP growth vary with the value of the cost share of capital assumed, although the parameter value does not change the results qualitatively (Table 1).

Figure 7.Growth Decomposition with Assumed Cost Share of Capital

Source: IMF staff estimates.

Table 1.TFP Growth Estimates, 1990–2011(Percent)
AlphaTotalOilNon-oilPrivatePublic
1990–2011
0.900.6-1.71.4-0.42.5
0.670.6-1.71.3-0.32.4
0.500.5-1.61.2-0.32.3
0.300.5-1.61.1-0.32.2
2005–2011
0.90-0.2-9.02.93.7-1.3
0.670.0-7.22.83.4-0.5
0.500.2-6.02.73.10.1
0.300.4-4.52.62.80.8
Source: IMF staff estimates.

TFP estimates from Solow residual estimation

6. The estimation approach allows α to be estimated for different sectors, yielding better estimates of TFP growth. The framework is also flexible enough to detect structural breaks in the dynamics of TFP growth. For instance, the result from assuming α = 0.67 shows a possible structural break of TFP growth around the year of 2005. To proceed, we start with the production function:

where Δyt=Δlog(YtLt),Δkt=Δlog(KtLt), and let Δ log(At) = γt + εt. In order to capture a possible structural break of TFP growth rate around year 2005, let γt = γ0 + γ1D2005, and D2005 is the post 2005 dummy indicator. Thus, now:

Δyt = γ0 + γ1D2005 + αΔkt + εt.

This specification is estimated for different sectors using data from 1990 to 2011; with the results presented in Table 2. Based on estimates of α, the assumption that α = 0.67 appears high for the economy overall, and it is on the high (low) side for oil (non-oil/private) sector, suggesting capital is more scarce (abundant) in the non-oil/private (oil) sector. The estimation also detects a structural break in TFP growth around 2005 as estimates of γ1 are both statistically and economically significant for all sectors except for the overall economy (the estimate for the oil sector is only significant at the 10 percent level).

Table 2.Solow Residual Estimates
TotalOilNon-oilPrivatePublic
y00.006230.003520.00724**-0.0219***0.0380***
(0.00706)(0.0237)(0.00267)(0.00339)(0.00600)
y1-0.00334-0.0614*0.0147**0.0579***-0.0439***
(0.00913)(0.0351)(0.00559)(0.00967)(0.00850)
α0.385**0.470***0.919***0.831***0.688***
(0.184)(0.155)(0.158)(0.107)(0.0958)
R-square0.1220.2390.7970.8220.641
Observations2222222222
Robust standard errors in parentheses*** p<0.01, ** p<0.05, * p<0.1Source: IMF staff estimates.

7. The results suggest that while there has been no TFP growth in the economy on aggregate (first column of Table 2), TFP growth has been strong in the non-oil sector since 2005 at almost 2.2 percent. The key driver is the private sector with TFP growth estimated at 3.6 percent a year after 2005. Figure 8 shows the sectoral contributions to growth using the respective estimates of α from Table 2. Again, however, how much of the break in 2005 is due to economic versus statistical issues is not clear.

Figure 8.Growth Decomposition with Estimated Cost Share of Capital

Source: IMF staff estimates.

8. Capital accumulation is the key driver of non-oil sector growth. The estimates of the cost share of capital are much higher for the non-oil sector than the oil sector, which reflects capital scarcity in the non oil sector and at the same time makes capital accumulation more effective at generating growth in the sector. Table 3 presents the average factor contribution to growth over the period 2000–11 using these estimates. The TFP contribution estimates for the oil sector are negative on average in both estimation methods. This could reflect the increasing difficulty of oil extraction as oil fields mature,4 or it could reflect the decision of Saudi Arabia to increase investment so as to have enough spare production capacity to ensure the global oil market remains well supplied in the event of unexpected demand or supply developments.

Table 3.Average Contribution to Real Output Growth, 2000–11(Percent)
TotalOilNon-oilPrivatePublic
Real output growth5.22.06.37.04.6
Alpha = 0.67
Labor1.60.91.61.81.3
Capital3.45.22.73.52.4
TFP0.2-4.11.91.70.9
Solow residual estimation
Labor2.91.50.40.91.2
Capital1.93.73.84.42.5
TFP0.3-3.12.11.70.9
Source: IMF staff estimates.

C. Estimates of Potential Output in Saudi Arabia

9. Robust growth of output in the non-oil sector is likely to continue going forward. Time series techniques are used to assess trend growth over the medium term of different sectors. These include filtering approaches that isolate high-frequency from low frequency components (Hodrick-Prescott, Baxter and King, and Christiano and Fitzgerald filters), and the Beveridge and Nelson decomposition. Estimates of growth rates from different methods are presented in Table 4. The average of these estimates is used to calculate trend growth. The private sector is expected to continue to be a key driver of non-oil growth with the highest growth rate of 6 percent over the projection period. Figure 9 shows the trend output growth dynamics over the medium term.

Table 4.Average Estimated Output Growth, 2012–18(Percent)
OilNon-oilPrivatePublic
HP1.45.96.44.2
BK1.55.46.03.8
CF1.55.66.13.9
BN1.64.75.43.2
Source: IMF staff estimates.

Figure 9.Trend Growth Projections, 2000–18

Source: IMF staff estimates.

10. Forecasts based on the growth accounting methodology suggest slightly stronger growth in the non-oil sector than the filtering approach results when taking into account the projected investment pattern during 2012–18. Using the average factor accumulation rate over the period of 2000–11, non-oil sector growth projections are calculated with TFP estimates from the growth accounting framework (Table 5). The estimates are slightly higher than the trended output projections suggested by the filtering approach. Using the factor accumulation forecast by staff over 2012–18, projected growth in the non-oil sector is similar (Table 5). This is because the key driver of growth, capital accumulation, is very close to the historical average over the forecast period.

Table 5.Non-oil Sector Growth Projections, 2013–18(Percent)
Non-oilPrivatePublic
Based on Average Factor Accumulation Rates (2000–11)
Capital4.65.33.6
Labor4.85.43.8
Projected sector growth
Alpha = 0.676.05.06.1
Solow residual6.98.93.0
Based on Forecast Accumulation Rates (2012–18)
Capital4.85.24.0
Labor12.52.52.5
Projected sector growth
Alpha = 0.675.34.05.9
Solow residual6.88.32.9
Source: IMF staff estimates.

11. The non-oil sector growth, which has surpassed potential in recent years driven by the expansionary fiscal policy, is likely to cool down and converge to potential over the medium term. The potential non-oil sector growth path between 2000 and 2011 is constructed using the long term factor accumulation rates and smoothed TFP growth rates under the growth accounting framework. As shown in Figure 10, actual non-oil growth exceeded its long run over the period of 2004–08, and resulted in a significant pick-up in inflation in 2008. After falling below potential in 2009, non-oil growth exceeded potential in 2010–11. This path, to a large extent, was driven by the growth in real fiscal expenditures since 2004. As fiscal spending moderates over the medium term, as expected by staff, non-oil growth will likely remain around its potential rate over the projected period.

Figure 10.Non Oil Sector Growth, 2000–18

Source: IMF staff estimates.

D. Conclusion

12. The recently revised GDP data show that productivity in the non-oil sector has been on the rise since 2005, largely driven by the private sector. Nevertheless, the drag from the oil and public sectors means there has been little productivity growth for the overall economy.

13. The analysis suggests that the strong growth of the non-oil sector is likely to continue over the medium term. One caveat to this is that a slowing in the pace of government spending could result in slower factor accumulation going forward.

14. Policies that aim at promoting potential output should target the private non-oil sector where resources appear to be allocated more efficiently. In addition, maintaining the growth momentum in the non-oil sector will be important for the authorities to raise Saudi employment in the economy.

References

    Solow, R. M.,1957, “Technical Change and Aggregate Production Function”, Review of Economics and Statistics, Vol. 39 (3), pp. 31220

    Tzimas, E., A.Georgakaki, C.Garcia Cortes, and S.D.Peteves,2005, “Enhanced Oil Recovery Using Carbon Dioxide in the European Energy SystemEuropean Commission Joint Research Center, Report EUR 21895 EN (Brussels: European Commission Joint Research Center).

    Hodrick, R, J., and E.Prescott,1981. “Post-War U.S. Business Cycles: An Empirical Investigation,Discussion Papers 451, Northwestern University, Center for Mathematical Studies in Economics and Management Science (Evanston, IL).

    Baxter, M., and R. G.King,1999. “Measuring Business Cycles: Approximate Band-Pass Filters For Economic Time Series,Review of Economics and Statistics, Vol. 81(4), pp. 575593.

    Christiano, L. and T.J.Fitzgerald,2003, “The band pass filter,International Economic Review, Vol. 44(2), pp. 43565.

    Beveridge, S., and C. R.Nelson,1981, “A new approach to decomposition of economic time series into permanent and transitory components with particular attention to measurement of the ‘business cycle,’Journal of Monetary Economics, Vol. 7, pp. 151174

Prepared by Haonan Qu.

Throughout this note, we divide the Saudi economy into three broad sectors: oil sector, public sector (government services and public utilities), and private sector (all other economic activities). The public sector and private sector constitute the non-oil sector in this context. Import duties are excluded for the purpose of the analysis.

Other Resources Citing This Publication