Book
Share
Chapter

20 A Structural Analysis of the Determinants of Inflation in the CEMAC Region

Author(s):
Andrew Berg, and Rafael Portillo
Published Date:
April 2018
Share
  • ShareShare
Show Summary Details
Author(s)
Rafael Portillo

1 Introduction

In this chapter1 we aim to identify the main determinants of inflation in the Communauté Economique et Monétaire de l’Afrique Centrale (CEMAC) and the transmission channels involved. The CEMAC is a monetary union comprising six countries in Central Africa (Cameroon, Congo, Gabon, Equatorial Guinea, the Central African Republic, and Chad), with monetary policy run at the regional level by the Banque des Etats de l’Afrique Centrale (BEAC), and with the local currency (the CFA franc) fixed to the euro.2 Our focus is on the region as a whole rather than on the individual countries.

We are particularly interested in the role of fiscal policy in inflation dynamics. Oil is the largest export sector in the CEMAC, and a large share of oil revenues accrues to the national governments.3 Fluctuations in oil revenues can therefore lead to sizeable changes in the fiscal stance, aggregate demand, and inflation. Movements in the latter variable also play an important role in external adjustment. This is because oil windfalls and the associated fiscal policy response affect the equilibrium real exchange rate, but under a fixed exchange rate this can only be achieved through changes in inflation. This dual nature of inflation is a key part of the macro policy challenge in pegs, and the CEMAC is no exception.

Our analysis consists of two parts. First, we use a semi-structural VAR analysis to identify the sources of inflation empirically. Second, we develop a dynamic stochastic general equilibrium model to analyse the channels through which oil-revenue-driven fiscal policy propagates in the region. We find that fiscal shocks—measured as changes in the non-oil fiscal stance—have been an important source of inflation, accounting for about 20 per cent of inflation volatility over the last ten years. In addition, passive monetary accommodation plays an important role in the propagation of fiscal shocks. This is not because of direct monetary financing or fiscal dominance more generally. Instead, fiscal expansions in commodity exporters are associated with improvements in the balance of payments, which also result in endogenous, unsterilized increases in reserve accumulation under the fixed exchange rate, and a corresponding monetary accommodation. By allowing inflation to increase, monetary policy helps deliver the real appreciation mentioned above.

We also use the model to study the implications of a more active monetary policy, in which case the central bank balance sheet would no longer expand endogenously in response to an oil-driven fiscal expansion. Under a peg, the latter policy is only possible if the capital account is closed, which is a strong but somewhat plausible assumption for the CEMAC region. An active monetary policy would be associated with greater (sterilized) reserve accumulation, which would help contain equilibrium appreciation pressures and therefore the pressures on inflation. The cost is that the private sector would be crowded out to create the necessary savings to support higher reserves. The policy lesson is that attempting to use monetary policy to contain inflation under a peg has important drawbacks, which further highlights the importance of prudent fiscal policy for macro and price stability in these regimes.

The chapter is organized as follows: Section 2 presents some visual evidence regarding the link between fiscal policy (and other variables) and inflation. Section 3 presents results from a VAR-based approach. Section 4 presents the DSGE model, while Section 5 presents model-based simulation results. Section 6 concludes.

2 Structural Determinants of Inflation in the Cemac Region: The Usual Suspects

The structural analysis of inflation requires the identification of the main shocks that may affect the price level, as well as the channels through which these shocks operate. A list of potential candidates includes fiscal policy, monetary policy, imported inflation, domestic supply shocks, and changes in regulated prices. While they are all likely to be important in accounting for price fluctuations at any given quarter, this chapter does not address the issue of regulated prices. Each of the remaining shocks is discussed below.

2.1 Fiscal Policy

As discussed above, oil-driven fiscal policy may be an important source of aggregate demand. This is especially the case if government spending is concentrated primarily on local goods and services. To identify the fiscal impulse in oil-rich economies requires that oil revenue be excluded from the measure of the fiscal stance. The resulting non-oil fiscal balance for the entire region, defined here as the sum of non-oil deficits over the sum of non-oil GDP for each country is presented in Figure 20.1 along with end of period inflation for the entire region.4

Figure 20.1.CEMAC: Inflation and Non-Oil Fiscal Deficits 1998–2008

There is a clear link between these two variables at annual frequency. However, there is the risk that the positive relationship may be driven by a third variable. For example, if the real depreciation of the CFA franc is positively correlated with inflation (which is the case, as will be shown later), then the co movement in Figure 20.1 could be driven by public expenditure items that are sensitive to exchange rate fluctuations, such as interest payments on external debt, or expenditure items concentrated on imported goods, such as capital expenditures. Since these items should not lead to an increase in demand pressures, the measure of the fiscal stance is further refined by removing these components.5Figure 20.2 shows the time series for this alternative measure.

Figure 20.2.CEMAC: Inflation and the Fiscal Stance 1998–2008

The co movement between the fiscal stance and inflation is now higher than before: the correlation is 0.83. In order to understand where this co-movement is coming from, we explore the link between the fiscal stance and other variables. First, the fiscal stance is also correlated with non-oil GDP growth (Figure 20.3), which suggests that the fiscal stance has a considerable impact on aggregate demand.

Figure 20.3.CEMAC: Non-Oil Growth and the Fiscal Stance 1998–2008

In addition, the region’s fiscal stance is also closely correlated with average money growth.6Figure 20.4 displays the co movement between the fiscal stance and money growth, measured both in terms of money and in terms of non-oil GDP.

Figure 20.4.CEMAC: Average Money Growth and the Fiscal Stance 1998–2008

Two features of the relationship are worth emphasizing. First, there is a strong co movement between money growth and the fiscal stance: the correlation is 0.83. Second, the magnitude of the changes in broad money, measured as a fraction of non-oil GDP, is similar to the magnitude of the fiscal stance. This relation does not operate through the standard concept of fiscal dominance, which would interpret such relations as driven by direct monetary financing of the deficit. Indeed, the comovement between the fiscal stance and growth in net credit to the government during this period is —0.37. Instead, this comovement is the result of higher government spending out of oil revenue, which is associated with greater reserve accumulation and monetary policy accommodation. This mechanism will be explored in our model.

2.2 Monetary Policy

In principle, under a fixed exchange rate regime (fixed to the euro in this case), monetary policy is not autonomous. Capital mobility implies that the domestic short-term interest rate is determined by the rate in the euro plus the risk premium associated with that country or group of countries. Moreover, the stock of money is determined by the liquidity needs of economic agents and the inflows of foreign capital to finance purchases of local assets and goods. In the CEMAC region, however, there is de facto limited mobility of capital, which provides some degree of independence for monetary policy and implies that the short-term interest rate may deviate from the euro rate, or more generally that monetary policy shocks may be a source of inflation pressures. In practice, however, monetary policy is mostly passive, in that the policy stance reflects endogenous changes in money demand and in the balance of payments, often as a result of fiscal policy developments. For this reason, we do not consider monetary policy shocks, i.e., exogenous changes in the policy stance, as a likely candidate for driving inflation. However, it also implies that endogenous changes in monetary policy can play an important role in inflation dynamics. Note that we will also consider the implications of a more active monetary policy with the use of the DSGE model.

2.3 Imported Inflation

Imported inflation is another likely candidate that can help account for fluctuations in prices. Figure 20.5 displays imported inflation, measured as an import-weighted index of inflation in the CEMAC’s main trade partners multiplied by their nominal exchange rate vis-à-vis the CFA franc:

Figure 20.5.CEMAC: Domestic and Imported Inflation 1996:1–2007:4

From Figure 20.5, imported inflation leads domestic inflation by two or three quarters. It can therefore help predict domestic prices at a relatively high frequency.

2.4 Domestic Supply Shocks

Finally, inflation displays large one time changes from one quarter to the next, as shown in Figure 20.6.

Figure 20.6.CEMAC: Domestic Quarterly Inflation 1996:1–2007:4

While some of these quarterly changes reflect the lower frequency fluctuations in the fiscal stance, part or most of this high-frequency variation is likely due to domestic supply shocks, such as changes in domestic agricultural production. Indeed, the low average inflation observed in 2007 is attributed in part to agricultural developments in Chad. These high-frequency changes could also reflect changes in regulated prices and/or measurement error, which is an important issue in the region. However, we will not dwell on those issues for the purposes of this analysis.

The previous analysis has indicated that there are three major likely candidates for understanding inflation developments in the region: fiscal policy shocks, shocks to imported inflation, and domestic supply shocks. It is important to identify the contribution that each one of these shocks has made to the overall variance of inflation. We estimate such contribution with a semi-structural VAR for the CEMAC economy, presented below. Secondly, we compare those results with a DSGE model, with an emphasis on the fiscal shock.

3 Assessment of Structural Shocks with a Semi-structural Var

In light of the evidence presented in Section 2, we estimate a vector auto regression (VAR) with two lags and five variables: imported inflation, the fiscal stance, non-oil GDP growth, money growth, and inflation. We make the following identifying assumptions. First, we assume that imported inflation is exogenous relative to the other variables. Second, we identify fiscal shocks by assuming that these respond contemporaneously to shocks in imported inflation only.7 Third, we interpret domestic supply shocks as unexpected changes in inflation that are orthogonal to contemporaneous movements in all other variables. This identification scheme seems plausible but, as with all VAR identification schemes, is not without drawbacks. For example, it may lead to a downward bias in the estimated contribution of supply shocks to overall inflation.8

We first present the variance decomposition of all variables by type of shocks in Table 20.1. Since the VAR is only semi-structural, the contribution of the three shocks need not add up to one. Imported inflation accounts for about 36 per cent of the volatility of inflation, while shocks to the fiscal stance account for about 20 per cent and domestic supply shocks explain an additional 16 per cent. Fiscal shocks also account for a sizeable fraction of the volatility of money growth, consistent with Figure 20.3. The variance decomposition only provides the contribution of fiscal shocks, rather than the contribution of the fiscal stance itself. Indeed, part of the contribution of higher import prices to inflation may be operating through an endogenous response in the fiscal stance: higher imported inflation, if it represents an appreciation of the dollar, may fuel an expansionary fiscal response since oil revenues are increasing in FCFA. The large role of imported inflation in the VAR may thus be explained in part by the fiscal expansion that may follow, and the share of money growth volatility that is accounted by the latter attests to that hypothesis.

Table 20.1.CEMAC: Variance Decomposition by Type of Structural Shock
Fiscal shocksShocks to imported inflationDomestic supply shocks
Share of total volatility (in per cent)
Inflation18.2336.2515.9
Money growth27.7157.350.42
Non-oil GDP growth30.5136.259.4
Fiscal stance37.2942.335.3
Imported inflation01000
Source: Staff calculations
Source: Staff calculations

Figure 20.7 presents the impulse response function for all variables (except imported inflation, which by assumption is exogenous) following a one percentage point increase in our measure of the fiscal stance, with the tenth and ninetieth percentiles of their distribution.9 Inflation increases by as much as 1.7 per cent after five quarters, while money growth increases by almost 7 per cent and non-oil GDP increases by about 1.5 per cent. These estimates are in line with the comovement we observed in previous graphs and suggest that the non-oil fiscal stance is an important source of aggregate demand in the region.

Figure 20.7.CEMAC: Impulse Response Functions Following a 1 Per Cent Increase in the Fiscal Stance, Based on VAR Estimates

We now describe a macroeconomic model that can help us rationalize the evidence just presented.

4 The Model

The DSGE model is a simplified version of the model in Chapter 12. The model has been extended by introducing habit formation in utility and backward-looking indexation in the Phillips curve. This section briefly presents the equations of the model in log-linearized form (i.e., in percentage deviations from the steady state), organized by economic agent.10

4.1 Consumers

Every quarter, the representative consumer chooses an inter-temporal path for consumption (Ct), demands financial assets and money, and makes labour supply decisions. Consumption decisions are given by the now-standard forward-looking IS equation:

where Rt is the short-term nominal interest rate and Etπt+1 is the expected quarterly inflation rate. The parameter h denotes the degree of habit formation in consumption, while σ is the relative risk aversion coefficient. Consumers also choose how to allocate consumption between three types of goods: non-traded goods and services (CtN), domestic agricultural products (CtNF), and imported goods (IMPt). Demand for each is as follows:

Demand depends on the prices of each type of good relative to the overall cpi: ptN,ptIMP,andptNF.11 The parameter χ refers to the price elasticity of demand. We assume that the law of one price holds, which implies ptIMP=st, where st (St) is the real (nominal) exchange rate (st = st-1 + St – St-1πt).

The allocation of consumption across goods leads to the following equation for inflation (as deviation from steady state inflation, which equals zero):

Imported inflation is set to zero for simplicity (πtIMP=0). Regarding financial assets, consumers choose between domestic and foreign assets (btandbt*), subject to a portfolio adjustment cost—measured in terms of changes in holdings of foreign assets—that prevents uncovered interest rate parity from holding in the model. This leads to the following relation between domestic interest rates Rt and the expected rate of depreciation of the nominal exchange rate (St+1St):

The portfolio adjustment cost, given by the coefficient ϕ, is a proxy for the degree of capital mobility in the model. Demand for real money balances is as follows:

Finally, labour supply (Lt) is given by a Frisch-type equation derived from utility maximization:

4.2 Firms

There are two local non-oil sectors in the economy: the non-traded sector and domestic agriculture. Firms in the non-traded sector produce goods using local labour: YtN=αLtN. Firms have some degree of market power and set the prices at which they sell their products. Price setting is also subject to nominal rigidities, as well as some degree of indexation to past prices, all of which leads to a hybrid Phillips curve for non-traded goods inflation πtN:

Non-traded inflation increases if real marginal costs increase relative to real prices in the sector. Real marginal costs are given by:

Finally, production of local agricultural goods is assumed as exogenous and set to zero YtNF=0. Non-oil GDP is given by the sum of these two sectors, weighted by their steady state shares:

4.3 The Government

To simplify, we assume that the government’s sole source of revenue comes from oil, which is modelled as an exogenous flow of foreign resources. In addition, we assume that all government spending is focused on the non-tradable sector. The government’s inter-temporal budget constraint is the following (θX refers to the steady state level of variable X in per cent on non-oil GDP):

An increase in oil revenue will lead to an increase in government spending (Gt), all else equal. In addition, a real appreciation of the currency (st ↓) and/or an increase in the relative price of non-tradable (ptN) will reduce the purchasing power of oil revenues and will lead to a decline in real government spending.

The variable bt denotes deviations in real debt from its steady state value. bct denotes the log-level of real debt that is held by private agents, on which the government pays real interest rate β-1 – 1 (at steady state).12 We assume that the government perfectly targets a constant level of real debt, which implies bt = 0.

Oil revenues follow an exogenous AR(1) process:

We also define the non-oil fiscal balance as a percentage of non-oil GDP as gy=ptN+GtYt.

4.4 The Central Bank

In real terms, changes in the central bank balance sheet are given by:

where bcbt refers to the central bank’s holdings of government debt and RR*t is the level of international reserves. Reserve accumulation is as follows:

Reserve accumulation depends on the degree of exchange rate targeting. In the case of a fixed exchange rate regime such as the CEMAC’s, ωs is calibrated to be infinitely large, which implies St=St1=S¯=0.

With regards to monetary policy, we assume that the size of the central bank balance sheet (and therefore the stance of policy to some extent) is determined by the endogenous accumulation of international reserves. This is implemented by setting the level of the central bank’s net domestic assets to be constant in nominal terms, which implies the following process for the real level of domestic assets (in logs):

Such an exogenous rule is consistent with monetary policy implementation in the CEMAC region, which as we discussed earlier is mostly passive.

4.5 Market Equilibrium Conditions

Closing the model requires ensuring that markets clear. There are three equilibrium conditions and one resource constraint (or external balance condition) that must hold at every period. First the labour market must clear:

Second, the non-traded goods market and the market for locally produced agricultural products must also clear:

Finally, external balance requires that imports be financed by: (i) oil revenues (OILt), (ii) drawing down on private sector foreign assets (bt*), or (iii) drawing down on international reserves (RRt*). This yields:

5 Calibration and Simulation Results

The model has twenty-four parameters, of which four are imposed by steady state identities (θbcbbOILIMP) Seven parameters (θGbcmRRN,n1,n2) are based on broad features of the CEMAC region. Calibration is presented in Table 20.2. The portfolio adjustment cost (ϕ) and the degree of exchange rate targeting via reserves (ωs) are set ‘infinitely high’ (105) to replicate a closed capital account and a fixed exchange rate regime, respectively. The country’s steady state private net foreign asset position (θb*) was set to zero for simplicity. The habit parameter (h) and the backward-looking term in the Phillips curve (ρπ) were chosen to generate plausible inflation dynamics in the model, while the choice of parameters (α,β,η,γ) follows standard practice in the DSGE literature. The risk aversion coefficient (σ), the elasticity of inflation to real marginal costs π), and the elasticity of substitution across types of goods (χ) reflect our priors that in developing countries: (i) aggregate demand is not very sensitive to changes in financial conditions, (ii) inflation is moderately sensitive to movements in output, and (iii) there is limited substitutability between imports and domestic goods. Finally the persistence of oil revenues OIL) is consistent with the prolonged, though ultimately temporary, nature of fluctuations in oil revenues.

Table 20.2Calibration
h0.90α0.6000θG0.15
σ5.00ρπ0.1000θOIL0.149
χ1.10κπ0.0500θB0.15
n10.35λN0.5000θBC0.1
n20.50ρOIL0.9000θBCB0.05
ϕ105β0.9900θIMP0.15
γ-1.50ωS105θM0.15
η0.50θb90.01θRR0.1

Figure 20.8 presents the impulse response functions in the model following a one percentage point increase in the fiscal stance, driven by higher oil revenues (utOIL>0). The paths are similar to those found in the VAR, except that the responses display less persistence and the growth rate of money is smaller. This relatively good match between data and model supports using the model to study the channels of transmission of fiscal policy and the interaction with monetary policy, which we do next.

Figure 20.8.Impulse Response Functions Following a 1 Per cent Increase in the Fiscal Stance, Model Simulations

5.1 The Impact of Fiscal Shocks in the DSGE

Here we elaborate on the impact of an increase in the non-oil fiscal deficit according to the model. When the government receives the additional oil income, and before the fiscal stance is changed, it is typically the case that the additional external revenue is deposited at the central bank. International reserves increase by the amount of the oil revenue, while net domestic assets of the central bank decrease, as government deposits are now higher.

As the government proceeds to spend this additional revenue, there is both a demand and a liquidity dimension to the spending. On the real side, provided it focuses on the non-traded sector, an increase in government spending will increase the demand for these goods, causing an increase in non-traded production and incipient inflationary pressures. The increase in the price of non-traded goods, i.e., the real appreciation, will lead consumers to substitute those goods for imports, which will help absorb part of the additional external revenues.

On the liquidity side, the drawing down of government deposits to finance the higher spending will lead a monetary injection. Under the assumption that monetary policy is passive, there is no offsetting operation to undo the effect on the monetary base. This monetary injection accommodates the increase in non-traded goods inflation described earlier. On the other hand, the increase in imports partially offsets the monetary expansion as it draws on the additional international reserves that were initially accumulated. Of course, ultimately what matters for the stance of policy is the level of nominal and real interest rates, which would be expected to decrease following the monetary expansion. However, interest rates are also affected by the increase in money demand. The stronger the increase in money demand, due to the expansion in nominal economic activity, the smaller the decrease in nominal and real interest rates.13 The latter effect helps dampen the inflationary pressures to some extent.

The end result from these various interactions is an expansion in activity, a temporarily higher inflation rate, a more appreciated real exchange, a decline in real interest rates though not necessarily nominal rates, an endogenous expansion in the size of the central bank balance sheet, and an increase in international reserves.

5.2 Active Versus Passive Monetary Policy

Is there a role for a more active monetary policy in this context? First, much depends on the degree of international capital mobility. Under our assumption of a perfectly closed capital account, there is scope for great monetary policy activism, since domestic interest rates are not pinned down by uncovered interest rate parity. We explore such activism by replacing the zero nominal growth rule for the central bank’s net domestic asset (bcbt = bcbt-1πt) with a rule specifying zero growth for nominal money balances:

The above rule implies that any monetary expansion resulting from the accumulation of reserves and drawing down of government deposits will be sterilized by selling part of the central bank’s stock of government bonds (bcbt ↓).14Figure 20.9 compares the macroeconomic effect of fiscal expansions under these two rules.

Figure 20.9.Impulse Response Functions, Passive Versus Active Monetary Policy, Model Simulations

By preventing the monetary base from expanding, the central bank is containing aggregate demand pressures in the economy, which is reflected in the smaller increase in inflation and smaller real appreciation, even though the fiscal expansion is broadly the same.15 A tighter control of aggregate demand leads to a decrease in private absorption and reduces the increase in imports, which implies a larger share of the additional oil revenue is accumulated as reserves. The latter is also consistent with a smaller real appreciation. The tighter monetary policy stance is reflected in higher real interest rates, compared with the passive case (at least over the first eight quarters).

5.3 Discussion

Does the above simulation suggest the central bank should be more active, or as active as is allowed by the degree of international capital mobility? Although it would appear that a tighter policy would be consistent with the price stability mandate, the reality is more complex. This is because an active monetary policy in this environment (peg and limited capital account) has real effects which go beyond the standard role of nominal aggregate demand management. In particular, the sterilized accumulation of reserves is affecting the degree of economy-wide absorption of the additional oil revenue, and comes at the cost of crowding out the private sector, in a context in which much needed additional external resources are becoming available to the country’s development. It is unlikely that the costs from preventing the private sector from accessing those resources would be warranted by the benefits from price stability. This further points to the dual role of inflation in countries with fixed exchange rates that we have stressed throughout this chapter.

The above discussion does not imply that there is little merit in accumulating reserves, or sovereign assets more generally, when there are external windfalls such as higher oil revenues. Nor does it imply that there is little to gain from greater macro and price stability in pegs. However, these objectives are best achieved by making fiscal policy less pro-cyclical, and by tying the accumulation of external assets to public savings (and explicit rules for the latter), as is the case for example with sovereign wealth and macro stabilization funds.16

6 Conclusion

The results presented in this chapter have highlighted the role of the fiscal stance in driving inflation in resource-rich countries with fixed exchange rates. The chapter has also presented a general equilibrium model to understand the channels through which fiscal policy affects inflation under a peg, and the role of passive versus active monetary policy rules.

From a policy perspective, the framework presented here could be useful for the BEAC when assessing its policy options. In particular, it could simulate alternative paths for the economy under alternative fiscal policy responses, as well as under different monetary policy arrangements.

References

    BergA.PortilloR.YangS. C. andZannaL. F. (2013). Public Investment in Resource-Abundant Developing Countries. IMF Economic Review61 (1) 92129.

    • Search Google Scholar
    • Export Citation

    BlanchardO.PerottiR. (2002). An Empirical Characterization of the Dynamic Effects of Changes in Government Spending and Taxes on Output. Quarterly Journal of Economics117 (4) 132968.

    • Search Google Scholar
    • Export Citation

    GaliJ. andGertlerM. (2007). Macroeconomic Modelling for Monetary Policy Evaluation. Journal of Economic Perspectives21 (4) 324.

    • Search Google Scholar
    • Export Citation

    GaliJ. andMonacelliT. (2005). Monetary Policy and Exchange Rate Volatility in a Small Open Economy. Review of Economic Studies7270734.

    • Search Google Scholar
    • Export Citation

    International Monetary Fund. (2008). Central African Economic and Monetary Community: Selected Issues Paper. IMF Country Report 08/254. Washington, DC: International Monetary Fund.

    • Search Google Scholar
    • Export Citation

    KilianL. (1998). Small Sample Confidence Intervals for Impulse Response Functions. Review of Economics and Statistics80 (2) 21830.

    • Search Google Scholar
    • Export Citation

    RameyV. andShapiroM. (1998). Proceedings from Carnegie-Rochester Conference Series on Public Policy 48: Costly Capital Reallocation and the Effects of Government Spending. North-Holland: NBER.

    • Search Google Scholar
    • Export Citation
1This chapter is a considerably revised version of IMF (2008).
2There are two monetary unions in Africa with their currency fixed to the euro, the second being l’Union Economique et Monétaire Ouest-Africaine (UEOMA), which comprises eight West-African countries: Benin, Burkina-Faso, Guinea-Bissau, Cote d’Ivoire, Mali, Niger, Senegal, and Togo.
3All countries in the CEMAC region export oil, with the exception of the Central African Republic (its main export commodity is diamonds). An important issue for the region is the expected depletion of oil reserves over the next two to three decades, which raises a number of macro challenges but is outside the scope of this chapter.
4Inflation numbers for the CEMAC region are constructed as a PPP-weighted geometric average of member countries’ inflation.
5There is a break in the series in 2002. This may reflect a change in the classification of spending or revenue items. However, the comovement between the series pre-break with the variables of interest (money growth, inflation, non-oil growth) is as strong—at least visually—as it is post-break, albeit a different level. For the purpose of the analysis, we correct the pre-break series with the difference in levels pre- and post-2002.
6We define average money growth as the average of the end of period money growth over four quarters.
7The use of VARs to identify exogenous changes in fiscal policy and trace their effects through the economy starts with Blanchard and Perotti (2002). Alternative measures rely on a narrative approach to identification, such as in Ramey and Shapiro (1998).
8The VAR is estimated with quarterly data for the period 1998 Q1–2007 Q4; the annual statistics for the fiscal stance and non-oil GDP growth are interpolated to create quarterly time series.
9Confidence intervals were derived using the bootstrap procedure described in Kilian (1998).
10For an introduction to DSGE models, see Gali and Gertler (2007) in the case of a closed economy and Gali and Monacelli (2005) in the case of a small open economy.
11Relative prices are related to inflation measures as follows: pti=pt1iπt+πti, for i = N, NF, IMP.
12The remaining stock of government debt is held by the central bank, such that: θbbt = θbcbct + θbcbbcbt:
13An increase in the nominal interest rate is also possible, though the increase in inflation makes real interest rate increases less likely.
14If the capital account is perfectly open, then efforts to sterilize the accumulation of reserves would result in large capital inflows, by adding incipient pressures on domestic interest rates, and further reserve accumulation. The central bank may stop sterilizing if the stock of reserves (and the associated sterilization costs) becomes too large.
15On the margin, the smaller real appreciation when monetary policy is active makes the (real) fiscal expansion larger.
16See for example, Chapter 12 and Berg et al. (2013).

    Other Resources Citing This Publication