This selected issues paper on Sudan was prepared by a staff team of the International Monetary Fund as background documentation for the periodic consultation with the member country. It is based on the information available at the time it was completed on September 7, 2012. The views expressed in this document are those of the staff team and do not necessarily reflect the views of the government of Sudan or the Executive Board of the IMF.

Abstract

This selected issues paper on Sudan was prepared by a staff team of the International Monetary Fund as background documentation for the periodic consultation with the member country. It is based on the information available at the time it was completed on September 7, 2012. The views expressed in this document are those of the staff team and do not necessarily reflect the views of the government of Sudan or the Executive Board of the IMF.

II. Sudan’s Inflation Problem: Some Lessons From The Past 30 Years1

Inflation in Sudan is a well-entrenched and persistent phenomenon. This analysis finds as key determinants: the exchange rate, reserve money, fiscal monetization, and wages. The exchange rate’s leading role implies that the dynamics of domestic inflation are heavily influenced by the external environment, highlighting the openness of Sudan’s economy, but also the focus on the exchange rate in the formation of inflationary expectations. Fiscal policy seems to have a limited direct effect but has an indirect impact via the government wage policy and the monetization of the budget deficit. Reserve money remains largely determinant in times of high inflation.

A. Background

1. During the 1980s and mid-1990s, Sudan suffered considerably from high inflation, including hyperinflation. Inflation was the result of weak economic and financial policies, including considerable financial (monetary and fiscal) expansion, severe import compression, extensive price and wage controls, rationing, high subsidies, and constraints on the private sector. These policies resulted in a large and active parallel market for goods and services and foreign exchange, and a surge of inflation which averaged 40 percent in the first half of the 80s only to accelerate thereafter, reaching 130 percent in 1991 and 140 percent in 1992.

2. Starting in 1991–92, in an effort to address the economic and financial situation, the authorities reduced their interference in the economy and opted for market-friendly policies. Almost all price and profit controls were abolished, and the prices of traded goods, with the exception of petroleum, sugar and wheat, reflected international prices. The foreign exchange rates and markets were unified, a move that required a devaluation of the exchange rate of more than 2000 percent. However, fiscal policy and monetary policy did not follow suit, with the result that inflationary pressure exceeded 150 percent by 1996.

3. In 1997–98, to address these deficiencies, Sudan embarked on a comprehensive and sustained structural reform program with the assistance of the Fund. The economy responded positively to these policies and reforms developed under the subsequent Staff-Monitored Programs (SMPs); inflation started to decline, dropping to 17 percent in 1998 and to 8 percent in 2000. However, double-digit annual inflation re-emerged in 2006 and persisted thereafter, reaching 19 percent in December 2011 and 37 percent in June 2012.

4. During the 1999–2010 period, Sudan’s macroeconomic performance was broadly satisfactory. Real growth averaged 6 percent and fiscal and external deficits remained modest at about 2 percent of GDP and 4 percent of GDP, respectively. However, inflation was the exception, remaining very high at an average of 13 percent.

5. This brief historical overview of inflation in Sudan during the past 30 years underscores that inflation is a well entrenched and persistent phenomenon. The central bank has only partially prevailed over it. Due to data limitations, this chapter will not cover the hyperinflation period, but instead will focus on the years 1998–2011. The latter period is important for two reasons: (i) It begins with the start of the oil era and ends with it; and (ii) with an average annual increase in the CPI of 13 percent, inflation emerges as the weakest link of a rather satisfactory macroeconomic performance.

6. This chapter is organized as follows: Section 1 presents an empirical analysis of inflation in Sudan, using quarterly data over the period 1998–2011. Section 2 discusses the calibration of a small model to forecast inflation. Section 3 discusses how reduced inflation could contribute to the stabilization of the macroeconomic framework and the successful transition to an economy less dependent on oil.

B. Main Features of Sudan’s Economy

Structure of the Economy

7. For the past 15 years since the discovery of oil, Sudan has been able to maintain a relatively diversified economy. During this period, the bulk of economic activities continued to be evenly distributed among three main sectors (Figure II.1):

  • Agriculture, whose contribution to GDP averaged about 32 percent;

  • Economic services, with a contribution slightly in excess of 30 percent of GDP; and

  • Mining, Manufacturing, Energy and Construction (MMEC), the fastest growing sector, owing to the discovery and expansion of the oil sector; with an average contribution of about 25 percent of GDP (Appendix II, Table 1).

Figure II.1.
Figure II.1.

GDP by Main Sectors

(In Percent of GDP at factor cost)

Citation: IMF Staff Country Reports 2012, 299; 10.5089/9781616355722.002.A002

Sources: Sudanese authorities; and staff estimates.

The remaining economic activities are constituted of social services (about 13 percent of GDP), including government services, whose share remained relatively stable at about 8 percent. Overall, while the oil sector’s value added was fairly modest (14 percent of GDP), its role in the financing of the country’s budget and BOP has been crucial (Appendix II, Table 2).

The CPI Basket

8. Sudan’s CPI basket contains 12 categories of goods and services, with tradable goods constituting 65 percent against 35 percent for the nontradable goods (Figure II.2; Appendix II, Table 3). The average annual inflation during the past 14 years was about 13 percent. This relatively high average is the result of a mix of double-digit inflation years (DDYs) with more modest or single digit inflation years (SDYs). The period 1998–11 comprises five SDYs with an average annual inflation of about 7 percent and nine DDYs with an average annual inflation of 16 percent (Figure II.3).

Figure II.2.
Figure II.2.

CPI Basket by Main Categories of Goods

Citation: IMF Staff Country Reports 2012, 299; 10.5089/9781616355722.002.A002

Sources: Sudanese authorities; and staff estimates.
Figure II.3.
Figure II.3.

Inflation Profile, 1997–2011

Citation: IMF Staff Country Reports 2012, 299; 10.5089/9781616355722.002.A002

1/ DDYs are the years 1998, 1999, 2002, 2007, 2008, 2009, 2010 and 2011.2/ SDYs are the years 2000, 2001, 2003, 2004, 2005.

Tradable versus nontradable goods

9. During the DDYs, high inflation was characterized by relatively high prices of both tradable and nontradable goods, with the highest rates registered for foodstuffs (tradables) and nonmarket services (nontradables) (Figure II. 4; Appendix II, Table 4).

Figure II.4.
Figure II.4.

Tradable Goods Inflation, 1997–2011

Citation: IMF Staff Country Reports 2012, 299; 10.5089/9781616355722.002.A002

1/ DDYs are the years 1998, 1999, 2002, 2007, 2008, 2009, 2010 and 2011.2/ SDYs are the years 2000, 2001, 2003, 2004, 2005.

10. The SDYs were characterized by a relatively low increase in the price of tradables (including food items) coupled with double-digit inflation on nonmarket nontradables (Figure II.5). The surge in the price of nonmarket nontradables in times of lower tradable goods inflation could suggest an underlying policy of loosening government control on these prices when inflationary pressures were abating.

Figure II.5.
Figure II.5.

Nontradable Goods Inflation 1997–2011

Citation: IMF Staff Country Reports 2012, 299; 10.5089/9781616355722.002.A002

1/ DDYs are the years 1998, 1999, 2002, 2007, 2008, 2009, 2010 and 2011.2/ SDYs are the years 2000, 2001, 2003, 2004, 2005.

Volatility and seasonality

11. Price volatility appears to be systemic, which would indicate an unstable economic environment. Seasonality, however, appears to be limited to foodstuffs (Appendix II, Table 5).

C. Modeling Inflation in Sudan

Data and Information Gaps

12. Collecting data for this analysis was difficult owing to large gaps in variables that have an impact on inflation.2

13. Modeling inflation in Sudan required making some assumptions regarding Sudan’s private consumption behavior. In particular, it has been assumed that Sudan’s households have:

  • a high and stable (over time and throughout the various population strata) propensity to consume (above 0.9);3 and

  • a utility function centered on: (i) first satisfying basic consumption requirements (tradables, and nontradables); and (ii) budget permitting, consuming secondary tradables and nontradables and/or savings.

These assumptions are in line with other low-income countries with moderate growth, which is the case for Sudan.

The Database

14. The model’s database includes quarterly series covering the period 1998/Q1–2011/Q4 for the following variables:

  • The Consumer Price Index;

  • The price index for tradable goods (TrCPI) derived from the CPI by aggregating the tradable categories of the CPI;

  • The price index for nontraded goods (NTrCPI) derived from the CPI by aggregating the nontradables categories of the CPI;

  • Reserve money;

  • Government current expenditure (GXC);

  • The exchange rate (SDGs per USD) $NER ; and

  • Household consumption in nominal terms (FCh) and real terms (FChQ), which have been derived by interpolating annual data and taking into account the CPI’s seasonal factors.

All variables have been converted into indices based in 1998/Q1: (1998/Q1 = 100).

15. In order to better understand the specificities of inflation dynamics during high and low inflation years, the data sample has been divided into two sub-samples covering the DDYs and SDYs.4

Key Trends During the Observation Period

16. The key trends during the observation period for the variables that will be used in the regressions are as follows:

  • Imported inflation as measured by the change in the import price in SDG terms is very close to the traded goods inflation, a strong indication that all traded goods, including nonexportables, are in line with international prices.

  • During low inflation years (SDYs), official reserves were high and the exchange rate appreciated, which allowed a substantial increase in imports of consumer goods in nominal terms (25 percent per year on average in SDG terms, almost twice that of private consumption) and in real terms (20 percent per year on average, more than three times higher than the real increase in private consumption) (Figure II.6; Table II.1).

  • On the contrary, during high inflation years (DDYs), reserves were low and the exchange rate weaker. While imported inflation measured in USD was fairly modest (more than 60 percent lower than that recorded during the low inflation years) it was almost four times higher in SDG terms, owing to the steep depreciation of the exchange rate. As a result, the increase in imports of consumer goods in real terms was reduced to an average of 8 percent per year, which remains, however, fairly high compared with the estimated 1½ percent real increase in final consumption.

  • Both central government current spending and wage bill grew much faster during low inflation years (46 percent and 31 percent, respectively) than during high inflation years (18 percent and 24 percent, respectively), which would indicate that these variables have only limited direct impact on inflation.

  • During both low and high inflation years, reserve money grew at the same high pace, around 25 percent per year on average. However, the fact that non-oil GDP real growth was much higher during SDYs (especially in the nonagricultural sector)5 than during DDYs would suggest that the demand for base money during the former was driven by economic growth and hence was less inflationary; also the velocity of base money in SDYs was higher than in DDYs. Conversely, with slower economic growth (especially in the nonagricultural sector), demand for base money during DDYs was mostly driven by inflation and inflationary pressures. That said, the absence of effect on inflation of the substantial increase in reserve money during low inflation years remains a puzzle that needs to be investigated further.

Figure II.6.
Figure II.6.

Inflation Developments, Quarterly 1998–2012, 1998/Q1 = 100

Citation: IMF Staff Country Reports 2012, 299; 10.5089/9781616355722.002.A002

Sources: CBS-Sudan
Table II.1.

Average Annual Changes of Key Determinants of Tradable Goods Inflation, (1998–2011)

article image

Source (GEE)

At constant SDG prices.

Using as a deflator the CPI.

D. The Main Model

17. In this analysis the inflation model for Sudan is centered on a set of three equations:

  • an equation explaining price developments for tradables;

  • an equation explaining price developments for nontradables; and

  • a third equation deriving total inflation as a function of tradable goods and nontradable goods inflation.

All the regressions have been done using the OLS methodology.

18. The model for tradable goods inflation is an auto-regressive model with one lag and three independent variables including current government spending X2,t; the exchange rate of the SDG against the U.S. dollar (with one lag) X3,t-1,6 and nominal private consumption (with one lag) X4,t-17:

Y1,t=0.41Y1,t-1+0.011X2,t+0.349X3,t-1+0.219X4,t-1R2=0.99(1)t-Stat(2.9)(2.01)(3.91)(4.32)

19. The specification of the model is appropriate and passes all the standard tests of statistical robustness (homoskedasticity, autoregressive conditional heteroskedasticity (ARCH), serial correlation, and unit root).8 Due to the lack of data on domestic production costs and household income, private consumption is used as a proxy variable for these costs which constitute the key component of domestic inflation. The assumption underlying this choice is that private consumption is highly correlated to household income (see footnote 2) which in Sudan constitutes the bulk of non-oil GDP. The role of this variable in equation (1) would then be to capture part of these costs and reflect them in tradable goods inflation.9 Overall, the domestic costs variable has the highest elasticity: a 10 percent increase in domestic costs, ceteris paribus, would generate an increase in tradable goods inflation of 3.8 percent, compared with 1.8 percent for the exchange rate and 0.6 percent for government spending.

20. The model for nontraded goods inflation has the same specification as the model for traded goods inflation, i.e., using level variables. It has three explanatory variables: tradable inflation (Y1,t), reserve money with one lag (X1,t-1), adjusted government current spending with one lag (X2,t-1), and the internal terms of trade (with one lag) (X6,t-1).

Y2,t=0.38Y1,t+0.11X1,t-1+0.03X2,t-1+0.74X6,t-1R2=0.99(2)t-Stat(5.3)(6.8)(5.2)(11.4)

21. The specification of the model is appropriate and passes the standard tests of statistical robustness (mainly homoskedasticity and serial correlation). The contribution of the independent variables to overall nontradables inflation is given by their respective elasticities: other things being equal, a 10 percent increase in lagged reserve money would raise nontradable goods inflation by 4 percent, compared with 3.1 percent for traded goods inflation, and 1.2 percent for government current expenditure. Also, an improvement in nontradables inflation during period t is positively correlated to the internal terms of trade during period (t–1) with an elasticity of 0.17.10 The small influence of government spending in the model for nontradable goods inflation appears as a confirmation that nontradable goods inflation is mainly driven by monetary policy. One could argue, however, that in this case, monetary policy (base money supply) is highly influenced by the size of the fiscal deficit, as evidenced by the fact that controlling for reserve money more than doubles the contribution of government spending.

22. The third equation of the model is a statistical equation summarizing the contribution of each of the tradable (Y1) and non-traded goods inflation (Y2) to total CPI inflation (Yt).

Yt=0.62Y1,t+0.39Y2,tR2=0.99(3)t-Stat(11.5)(12.7)

It is worth noting that the coefficients on Y1 and Y2 are somewhat close to their respective weights in the 2007 CPI basket.

E. The SDY Sub Model

23. The regression related to tradable goods inflation gives the following equation:

Y1,t=0.17X1,t+0.82X3,tR2=0.68(1-b)t-Stat(8.9)(26.4)

The specification of the model is different for that of the main mode and indicates that in time of low inflation, only reserve money X1t and the exchange rate X3,t emerge in the regression as significant independent variables; government expenditure does seem to have had no effect on tradable goods inflation despite the fact that during this period it had a relatively high average annual change (Table II.2). The estimated model underscores that during the SDYs, the appreciation of the exchange rate helped to contain tradable goods inflation which at the same time was being boosted by a lax monetary policy. This result should, however, be interpreted with caution, given that during the estimation period the official exchange rate appreciated only about 2 percent per year whereas reserve money was increasing at a pace of 27 percent per year on average. More specifically, according to the model a 10 percent appreciation/depreciation of the exchange rate, ceteris paribus, results in a 6.3 percent increase/drop in tradable goods inflation, compared with an increase of 3.7 percent for a 10 percent increase in reserve money.11

Table II.2.

Sudan: Inflation Model Elasticities

article image

24. The regression related to nontradable goods inflation gives the following model:

Y2,t=0.17X2,t+0.82Y1,tR2=0.86(2-b)t-Stat(7.1)(16.4)

25. Only government current spending and tradable goods inflation emerge in the regression as significant independent variables. Under this model, an increase in the price of tradables of 10 percent would result in an increase in the price of nontradables of about 6.3 percent compared with 3.7 percent for a 10 percent change in government spending.

26. As for the main model, the third equation of the model summarizes the respective contribution of tradable (Y1) and nontraded goods inflation (Y2) to total CPI inflation (Yt).

Yt=0.62Y1,t+0.39Y2,tR2=0.98(3-b)t-Stat(11.5)(12.7)

27. It is worth noting that the coefficients on Y1 and Y2 are close to their respective weights in the 2007 CPI baskets (0.65 and 0.35) than those of the main model.

F. The DDY Sub Model

28. The model for tradable goods is statistically robust with statistically significant parameters for all independent variables.12 The equation is as follows:

Y1,t=+0.13X1,t+0.24X2,t+0.34X3,t+0.27X4,tR2=0.99(1-c)t-Stat(2.3)(6.9)(6.4)(2.6)

The model indicates that in time of high inflation, the key macro variables that have an impact on inflation of tradable goods include reserve money, adjusted government current spending, the exchange rate, and private consumption as a proxy to domestic costs. In terms of elasticities, the exchange rate is second to domestic costs and is followed by adjusted government spending and reserve money. More specifically, a 10 percent increase in domestic costs, ceteris paribus, generates an increase in tradable goods inflation of 3.1 percent, compared with 2.4 percent for a 10 percent depreciation of the exchange rate, and 2.3 and 2.2 percent for adjusted government spending and reserve money, respectively. The elasticity of the exchange rate is close to the estimated import content of private consumption (about 20 percent) and confirms the important role of the curb exchange rate which has been used in the regression as of 2008. It is worth mentioning however that due to the cash nature of Sudan’s economy, the overall effect of the exchange rate is even higher when combined with its effect on demand for base money.

29. The regression related to nontradable goods inflation gives the following model:

Y2,t=0.33X1,t+0.17X2,t+0.67X6,t-1R2=0.98(2-a)t-Stat(13.7)(3.4)(11.5)

30. The model is statistically robust: it passes the standard tests (homoskedasticity and serial correlation), and the parameters of the independent variables are statistically significant. While both reserve money and government emerge as having a noticeable effect on nontradable goods inflation, it is reserve money that is revealed as the dominant independent variable with the highest elasticity. The third variable—the lagged internal terms of trade—plays the role of a memory variable in the nontradables inflation process, reflecting the structural changes in the relative price of nontradables in terms of tradables. More specifically, everything else being equal, a 10 increase in reserve money would result in an increase in the price of nontradables of 6.5 percent compared with 1.8 percent for a 10 percent increase in adjusted government current spending. The inflation component related to the internal terms of trade works as follows: a 10 percent increase in the internal terms of trade during (t–1) would generate a carry-forward to period t of a 1.7 percent of nontradables inflation.13

31. The third equation of the model is as follows:

Yt=0.67Y1,t+0.39Y2,tR2=0.98(3-c)t-Stat(15.1)(8.6)

32. It is worth noting that the coefficients on Y1 and Y2 are closer from their respective weights in the 2007 CPI baskets (0.65 and 0.35) than those of the main model.

G. Discussion of the Results

33. The results of the various models are recapped in the table below, which summarizes the elasticities associated with a 10 percent increase in the main independent variables used in the various regressions. Given their important policy implications, these results raise the following key questions that would need to be addressed:

  • What is the rationale for the apparent reduced impact of government spending on inflation?

  • Why does reserve money appear to have a more active role than the other variables, especially during high inflation years?

  • Why does the exchange rate play a role that goes beyond its indirect impact through imported goods?

34. The reduced role of government spending could be explained by the fact that: (i) part of this spending is sterilized through the balance of payments; and (ii) only the wage bill could have a direct effect on prices, which could only be limited given that the federal government wage bill represents about 7 percent of the total household income.14 This unexpected result15 does not seem to be in line with the accepted assumption that the discovery of oil in Sudan in the second half of the 1990s heightened fiscal dominance and limited the Central Bank of Sudan’s (CBOS) room to maneuver in fighting inflation.

35. However, the fact that the regressions have not revealed any form of government spending as a significant independent variable does not necessarily reject the existence of indirect correlation between government spending and consumer inflation. In particular, it is worth noting that there are two key indirect channels through which fiscal policy can affect inflation, namely: (i) The wage channel which goes through private consumption, a significant independent variable in both the tradable goods and non-tradable goods inflation models;16 and (ii) the monetary financing of the budget which goes through reserve money, whose effect is amplified by the cash nature of the Sudanese economy.

36. With respect to reserve money, almost all the estimated models indicate that inflation responds significantly to monetary stimulus when inflation has reached a relatively high level. It remains, however, a puzzle that during SDYs the substantial increase in reserve money (21 percent per year on average) did not have any marked effect on prices.17

37. The emergence of the exchange rate as a key determinant of tradable goods inflation confirms that in Sudan prices of tradables are broadly in line with international prices. This conclusion, which applies to both exportables and nonexportables, implies also that Sudan’s internal inflation dynamics for these goods (including production costs) are heavily influenced by the exchange rate and external prices. The large influence of the curb market exchange rate on the tradable goods CPI was quite evident during the past two years, as indicated by the statistical robustness of the regression of the latter on the former and the high correlation between the two variables.18

Performance of the model

38. The model is composed of a main model (MM) that has been estimated over the period 1998 Q1–2011Q4, and two sub-models SM1 and SM2 relative to the high and low inflation periods, respectively.

  • Performance of MM is broadly satisfactory: it is stable with a relatively strong analytical capability of developments of inflation in Sudan in the last 15 years. Its predictive capacity, however, is impaired by the fact that the estimated coefficients on the independent variables are associated with a rather nonhomogenous period of wide variations in inflation and policies.

  • Performance of the sub-model DDYM relative to the high inflation years is greater than that of MM, as the coefficients on the independent variables are associated with more homogenous macroeconomic conditions. In particular, this model appears relevant to the current situation in Sudan and could appropriately be used for projection purposes.

  • The model SDYM relative to low inflation years has been estimated just for illustrative purpose, given that its smaller underlying data sample strongly limited its statistical robustness. Eventually, it has not revealed any particular feature associated with low inflation years that could be used for policy purposes.

H. Conclusion and Policy Recommendations

39. Our statistical and econometric analysis of inflation in Sudan has underscored that its key determinants are reserve money, the exchange rate, fiscal monetization, and wage policy. While fiscal policy seems to have a limited direct effect, its indirect impact through the government’s wage policy, and its presumed impact on national wages and the monetization of the budget deficit, are substantial. Inflation in Sudan is also characterized by a domestic cost dynamic largely influenced by the external environment, which underscores the open nature of Sudan’s economy.

40. Over the past 14 years high inflation was contained thanks to the oil windfall, which provided the necessary financing for both the budget and the BOP and allowed an overall improvement in living standards. Policymakers were not under pressure to solve the inflation problem as there was no need and/or pressure to improve competitiveness and enhance import substitution, or to expand the tax base and improve revenue collection: the oil windfall was doing the job.

41. However, following South Sudan’s secession, Sudan entered a new era which requires changing course to adapt to its reduced economic and financial potential. To do so, Sudan has to rely on adjustment measures and endogenous resources to finance its budgetary and resource requirements. In this context, keeping inflation in check appears as the key policy target for stabilizing the macroeconomy, containing labor costs, and eventually enhancing competitiveness and improving the external current account.

42. The reform program that was adopted by the government in June 2012 is a first positive step towards addressing the adverse effects of South Sudan’s secession. Consolidating public finances, rebalancing the country’s external accounts, and bringing inflation under control will necessitate a strong and determined implementation of the government reforms and a diligent continuation of the adjustment efforts over the medium term.

43. In particular, reining in inflation will require a fruitful and close cooperation between the central bank and the ministry of finance. Based on the outcome of the above analysis two out of the three key determinants of inflation in Sudan (money supply and exchange rate) are under the control of the central bank, while the third (wage policy) is the responsibility of the ministry of finance. This outcome emphasizes the need to enhance the role of the central bank by, among other measures, increasing its independence.

Appendix II. Tables

Appendix II. Table 1.

GDP by Main Sectors

article image
Source: Central Bureau of Statistics; and Staff estimates.
Appendix II. Table 2.

Oil Sector’s Fiscal and External Predominance

(In percentage)

article image
Sources: CBOS; and staff estimates.
Appendix II. Table 3.

Sudan: CPI Basket by main Categories

article image
Source: Central Bureau of Statistics
Appendix II. Table 4.

Inflation profile, 1997-2011

(In percentage)

article image
Source: Central Bureau of Statistics

DDYs are the years 1998, 1999, 2002, 2007, 2008, 2009, 2010 and 2011.

SDYs are the years 2000, 2001, 2003, 2004, 2005, and 2007.

Appendix II. Table 5.

Inflation in Sudan: Dispersion Parameters

Low inflation Years VS High Inflation Years, Monthly

article image
1

Prepared by Ramdane Abdoun.

2

Quarterly fiscal data; non-oil GDP (income approach), economy’s average wage, average size of and wage in the civil service, employment in the nonagricultural sector; employment in the agricultural sector; non-monetary private consumption; quarterly import volume and price indices; quarterly non-oil export volume and prices indices; quarterly agricultural production indices (volumes and prices).

3

The propensity to consume is used to derive the households’ disposable income and savings.

4

In case there are two nonsuccessive years within the same sub-sample, a chain coefficient is used to connect them. Assuming that the years t and (t+2) are part of the DDYs sub-sample, for a given variable X, chaining the first quarter of year (t+2) to the last quarter of year t would consist in (i) chaining the first month of year (t+2) to the 12th month of year t using the change in percent between the first month of (t+2) and the 12th month of year (t+1); (ii) reconstituting the remaining months of year (t+2) using as a base the revised first month of year (t+2); and (iii) deriving the new quarters of year (t+2) by consolidating the revised months.

5

Although outside the purpose and scope of this chapter, it is worth mentioning the negative effect of inflation on non-oil growth, particularly in the nonagricultural sector.

6

Both the NEER and nominal exchange rate against the U.S. dollar have emerged as statistically significant regressors. The latter has been chosen because it is more a policy variable than the former and easy to project.

7

In the regression, this variable has been used as a proxy for households’ disposable income.

8

The small coefficient on government spending reflects the considerable increase in the government spending index (1998q=1100) during the observation period: At end-2011 it reached about 2,200 compared with some 240 for the exchange rate, 750 for the proxy for domestic costs and about 430 for the tradable goods CPI.

9

Excluding a domestic cost variable from the regression because of data unavailability could affect the reliability of the coefficients on the remaining variables. Using a proxy variable for an unavailable explanatory variable could mitigate the omitted variable’s bias and enhances the quality of the estimators of the coefficients on the other observed variables. (On using proxy variables, see Econometrics (J.M. Wooldridge).

10

The Itot variable has been chosen in place of the lagged nontradables inflation for which the null hypothesis of a unit root could not be rejected at a satisfactory significance level.

11

The modest size of the sample used for the regression and the low correlation coefficient call for caution when interpreting these results.

12

Contrary to the main model, in the DD model the lagged tradable inflation variable does not emerge with a statistically significant coefficient.

13

For example during (t–1) nontradable and tradable prices increased by 22 percent and 20 percent respectively, then a 1.7 percent price increase in nontradables is carried forward to period t.

14

Assuming that a Sudan household’s average propensity to consume is in the order of 90–95 percent, we can derive that compensation of employees in Sudan was about SDG 115 billion in 2010, about 11 times higher than the wage bill of the civil service of the general government (i.e., including the states).

15

There is a possibility that this is due to limitations of the fiscal data. Indeed, contrary to the monetary and exchange rate data, fiscal data are not compiled on a quarterly basis by the authorities. Thus, for the purpose of this note, these data had to be estimated using ad-hoc quarterly execution rates of the federal budget.

16

In fact, the regression of private consumption on the central government wage bill reveals a very high dependence of the former on the latter, which is consistent with the widely recognized role of the government as the wage setter in the economy. Based on available information, the government does not interfere in wage setting in the private sector, except for the requirement of a minimum wage level. In practice, the minimum wage is a key benchmark for the salary scale of the public sector and used as a base for wage negotiations in the private sector.

17

Part of the explanation could be that the old CPI (based in 1996) could not capture the full intensity of inflation during the SDYs.

18

The regression is based on monthly data covering the period January 2010–December 2011. The data on the curb market exchange are monthly and date back to January 2010. While there are no data on the curb market exchange rate prior to 2010, there are indications that the premium was minor, in the order of 3 percent or less.

Sudan: Selected Issues Paper
Author: International Monetary Fund. Middle East and Central Asia Dept.