Economic Growth, Inflation and Wage Growth: Experience from a Developing Country

Transcription

1 Business and Manageent Research Vol., No. ; 0 Econoic Growth, Inflation and Wage Growth: Experience fro a Developing Countr Shahra Fattahi (Corresponding author) Departent of Econoics Facult of Social Sciences, Razi Universit Shahid Beheshti Blvd, Keranshah, Iran Tel: Eail: sh_fatahi@ahoo.co Received: Januar 7, 0 Accepted: Ma 6, 0 Online Published: June 8, 0 doi:0.50/br.vnp88 URL: Abstract The Iranian econo has been subject to a nuber of critical events over the past five decades including the 979 revolution, the eight-ear war with Iraq, volatilit in global oil prices, and the 99 balance of paent crisis. These events plus governent controls of the ajor parts of the econo have substantiall changed the behavior of the acroeconoic variables. This research tries to exaine the behavior of the ke acroeconoic variables including econoic growth, inflation and wage growth over the period Based on the estiated ulti equation odel and appling the three scenarios, the econoic growth is predicted to oscillate between.8% and 6.98%. Furtherore, it is forecasted that the inflation rate and the wage growth rate will be increased. Kewords: Econoic Growth, Inflation, Forecast, Developing Countr. Introduction This paper tries to exaine the behavior of the ke acroeconoic variables including econoic growth, inflation and wage growth in the Iranian econo over the period The Iranian econo is oil-reliant so that an change in oil price can directl affect all econoic sectors. It should be noted that Iran ranks second in the world in natural gas reserves and third in oil reserves. It is also OPEC s second largest oil exporter. The Iranian econo has experienced a relativel high inflation averaging about 5 percent per ear. The inflation rate has even been ore than 0 percent on average after the 97 oil crisis. Another ajor proble is high uneploent in Iran. The uneploent rate increased fro.8 percent in 959 to 8.7 percent in 97 and this increase has continued so that it peaked at. percent in 989. Uneploent has been on average ore than 9. percent for the period High inflation along with high uneploent, which is referred to as stagflation, has been the ajor concerns in Iran s econo. Although the oil shock is the source of stagflation in developed countries, the assive currenc depreciation is the ain factor for the case of Iran (Bahani-Oskooee, 995). In order to reove these probles, the governent has taken soe actions including trade liberalization, tax refor, exchange rate unification, subsid refor and financial sector refor. However, little progress has been ade in the area of privatization. Oil revenues have been one of the ain sources of one creation, fueled b governent spending. Another iportant factor for increasing liquidit (M) is the large nuber of governent-controlled enterprises which increase the budget deficit through borrowing fro the Central Bank, and thus increase the onetar base. Mone suppl growth has been 5.7 percent on average for the period , whereas real GNP growth recorded onl on average 6. percent during the sae period. Furtherore, one suppl has becoe 66 fold while real GNP recorded onl a fold increase during the sae period. As the sae tie population has been growing, resulting in an increasing deand which adds to inflation pressure. The rest of this paper is organized as follows. Section presents the odel to be estiated and describes all the data used. Section reports the epirical results of the stud. Finall, section concludes.. A ulti-equation odel The ulti-equation odel to be estiated consists of three equations. The first equation is the wage equation: Published b Sciedu Press 88 ISSN E-ISSN 97-60X

2 Business and Manageent Research Vol., No. ; 0 e W 0 U OG () 0 0 where W is noinal wage growth rate, U uneploent rate, OG output gap inflation rate, and e expected inflation rate. Equation () is the expectations-augented Phillips curve. The output gap is defined as the percentage deviation of real GNP fro its long-ter trend (derived b the Hodrick-Prescott filter) OG * * * where is real GNP and is the potential output of the econo. The uneploent rate (U) is a prox for excess suppl of labor. An increase in U will result in a decrease in W so that one expects 0. Since is a prox for the deand for labor and there is a positive relationship between and OG, one would expect 0. The increases in wages caused b increases in prices are given b the equation W MP L. P, where MP L is the arginal product of labor. The role of price inflation and expected price inflation in the deterination of wages has been ephasized b an researchers (Lahiri, 98; Chen and Flaschel, 006). In a copetitive econo, the expected inflation should have a coefficient equal to unit while in noncopetitive situations it lies between zero and unit depending on the strength of unions and other noncopetitive eleents in the bargaining process. (Turnovsk and Wachter, 97). Furtherore, the expected inflation coefficient will be unit onl if workers can full account for the aount of expected inflation in their current wage settleents. According to the natural rate hpothesis, there is no wa for the governent to keep the uneploent rate peranentl below the natural rate. Therefore, there is no long-run trade-off between inflation and uneploent, which iplies that the coefficient of expected inflation will be unit. Equation () is the aggregate deand function 0 ( M / P) g c D () 0 0 where M / P is real one ( M broad one and P the iplicit price deflator), g real governent expenditure, c real consuption, and D du variable. This function can be derived as a solution of IS-LM relationships, with the lagged value of c introduced to ake the iplied consuption function of peranent-incoe hpothesis and the du variable included to capture the effect of the 979 revolution on production. All variables in equation () except the du variable will be eploed in the growth rates. Equation () deterines price level changes as 0 M X T () 0 0 where X is the labor productivit easured as the ratio of real GNP to eploent iport inflation, and T is the trend variable. The variables M and X will be used in the growth rates. Bahani-Oskooee (995), Liu and Adedeji (000), Valadkhani (006) and Bonato (007) have supported one suppl growth as one of the ain deterinants of inflation in Iran. It should be noted that wages are not the cause of inflation in Iran because labor unions do not have enough power and pla no deterining role in the econo. Since all equations are over-identified, appling the ordinar least square (OLS) ethod to the equations will be inappropriate. One solution to this proble is to use two-stage least square procedure (SLS). Table displas suar statistics of the data including the ean, the axiu, the iniu, and the standard deviation for the period The data are annuall and are collected fro the Central Bank of Iran. Before running the regressions, we exained whether the data are stationar. The epirical results of the Augented Dick-Fuller test indicated that all the variables eploed are stationar, and thus this issue helps us to avoid the proble of the spurious relationships.. Epirical results. Estiation In order to estiate the ulti equation odel paraeters, the process of expectation foration should be deterined. An adaptive learning odel, which is a ore realistic wa of odeling expectation foration, will be eploed Published b Sciedu Press 89 ISSN E-ISSN 97-60X

3 Business and Manageent Research Vol., No. ; 0 (Evans and Honkapohja, 00; Orphanides and Willias, 00; Basdevant, 005). Neither backward-looking odels (for instance, adaptive expectations odel) and nor odels with rational expectations are reasonable and realistic because the forer assues that historical econoetric relationships are invariant to changes in the econoic polic and that agents do not react to ssteatic istakes the ade while the latter assues that agents full know the structure of the econo and for their expectations accordingl. Since private agents are not endowed with a priori knowledge of inflation behavior, the ust learn about it over tie as new data becoes available (Sargent, 999). Although uch research has been done on adaptive learning odels in recent ears, the are largel theoretical (Bernanke, 00). In this stud, expected inflation is generated through a least square learning rule and then expected inflation is analzed in the augented Phillips curve equation. An econoetric tool to stud learning is the Kalan filter which can be used to estiate tie-varing econoic relationships. The learning approach assues that the agents expectations of inflation are on average correct but a liited set of inforation is used. It is assued that agents forecast inflation using t and t through recursive least squares (RLS) t t+ b, t b, t t b, t t t where t t+ is the inflation rate in the next period expected b the agents at tie t. Agents forecast inflation in the next period b updating the paraeters period b period. The process of updating is based on RLS as follows (Bullard 99, Sargent 999, Evans and Honkapohja 00) B B ' t R t t t X t( t X t B t ) ' Rt R t t ( XtXt R t ) ' where Bt ( b, t, b, t, b, t ) and Xt ( t, t ). The equations above correspond to the following state space odel: t t+ b, t b, t t b, t t t bit b it vit, The expected inflation is coputed as the predicted value for t t+. The results of the estiated ulti equation odel are as follows (p-value in parenthesis): W U 0.7OG e R 0.55 DW.68 (0.00) (0.000) (0.00) (0.000) (0.000) ( M / P) 0.g 0.9D R 0.5 DW.90 (0.96) (0.00) (0.00) (0.0000).7 0.9M 0.6 X T R 0.7 DW.89 (0.6) (0.0) (0.0) (0.0000) (0.00) Since there was evidence of autocorrelation in the wage equation, the estiated function was corrected for this proble. The result fro diagnostic checking show the odel does not suffer fro heteroscedasticit. In the incoe equation, lagged real consuption was excluded fro specification because this variable was not significant. All coefficients of the explanator variables in the ulti equation odel have the expected signs and are statisticall significant. In the wage equation, a higher uneploent rate is associated with a lower wage growth rate while higher output gap increases wage growth rate. The expected inflation is the ost iportant deterinant of changes in wages. The size of the coefficient of expected inflation (0.60) is relativel large, even larger than the coefficient of actual inflation (0.58). It eans that the wage growth rate is ore sensitive to the expected inflation than the actual inflation. Inflation expectations are ver unstable in Iran s econo because the Central Bank is unable to adhere to an inflation target in practice. Thus, inflation expectations are not well-anchored and an oil price increase, which sees apparentl to be a favorable shock, results in one creation, fueled b governent spending out of oil revenues, and inflation. Published b Sciedu Press 90 ISSN E-ISSN 97-60X

4 Business and Manageent Research Vol., No. ; 0 In the incoe equation, real growth rate of M and real growth rate of governent expenditure are positivel associated with econoic growth while the production has decreased due to the 979 revolution. In the last equation, the iport price growth rate is the ost iportant deterinant of changes in prices. The trend variable and noinal growth rate of M have alost the sae positive effect on inflation rate, 0.5 and 0.9 respectivel. Finall, the coefficient of the labor productivit is negative as expected.. Polic Scenarios Three scenarios are ipleented with different assuptions nael Scenario I (baseline scenario), Scenario II and Scenario III. In Scenario I, it assued that long run trends in endogenous variables as well as exogenous variables will continue to hold over the next five ears. In this case, a tie series ethod is ipleented to produce the projections for future values of the all variables. A collection of ARMA (p, q) odels, for different orders of p and q, have been estiated and then the best odel was selected according to the Akaike inforation criterion (AIC) and the Schwarz inforation criterion (SIC). The results indicated that the ARMA (, ) odel does a reasonabl good job for ost variables. Table shows the averages of the actual values over the period and the averages of predicted values over the period It is of special interest to exaine the behavior of the three endogenous variables (i.e. w, and ). While the average wage growth rate (w) was 7.9% over the last 50 ears, it is forecasted to rise to 9.5% over the period in the baseline scenario. The average inflation rate ( ) is predicted to be 8.% over the next 5 ears. However, the average econoic growth rate () is expected to increase slightl over the prediction period. It will increase fro 6.% to 6.98%. Scenario II takes into account the long run trends onl in exogenous variables in a different wa. In doing so, first, averages of the actual values of the exogenous variables are calculated over the period Then averages of predicted values of the endogenous variables can be obtained b substituting averages of the actual values of the exogenous variables into the estiated wage, incoe and price equations. Scenario III is siilar to Scenario II but the recent trends in exogenous variables will be considered. In fact, averages of the actual values of the exogenous variables are calculated over the period The results of the predicted endogenous variables under three scenarios have been presented in Table. Under all three scenarios, it is forecasted that the inflation rate and the wage growth rate will be increased over the period 008-0, as copared to the averages of corresponding values, for instance 5.% and 7.9% respectivel. Furtherore, ore inflation and higher wage growth are expected if the subsid refor plan is ipleented. The econoic growth is also predicted to oscillate between.8% and 6.98%.. Conclusions The stud exained the behavior of the ke acroeconoic variables including wage growth, econoic growth, and inflation using a ulti equation odel. The results indicated that the actual inflation and the expected inflation are positivel associated with wage growth rate. Furtherore, the inflation expectations pla a ajor role in the deterination of the wages. Econoic growth has been 6.% on average over the period However, this growth in Iran has not been sustainable. Furtherore, it should be noted that one suppl has becoe 66 fold while real GNP recorded onl a fold increase during the sae period. The Iranian econo has experienced a relativel high inflation with an average inflation rate of about 5 percent over the period It has even been ore than 0 percent on average after the 97 oil crisis. The iport price growth rate has been the ost iportant deterinant of changes in prices. The trend variable and noinal growth rate of M have alost the sae positive effect on inflation rate. It should be noted that oil revenues have been one of the ain sources of one creation, fueled b governent spending. Another iportant factor for increasing liquidit (M) is the large nuber of governent-controlled enterprises which increase the budget deficit through borrowing fro the Central Bank, and thus increase the onetar base. Mone suppl growth has been 5.7 percent on average for the period Finall, three scenarios were ipleented with different assuptions. Under all three scenarios, it is forecasted that the inflation rate and the wage growth rate will be increased over the period The econoic growth is also predicted to oscillate between.8% and 6.98%. References Bahani-Oskooee, M. (995). Source of Inflation in Post-Revolutionar Iran. International Econoic Journal. 9(). Published b Sciedu Press 9 ISSN E-ISSN 97-60X

5 Business and Manageent Research Vol., No. ; Basdevant, O. (005). Learning Process and Rational Expectations: An Analsis Using a Sall. Macroeconoic Model for New Zealand. Econoic. Modelling. (6) Bernanke B. S. (00). A Perspective on Inflation Targeting. [Online] Available: (March 5, 00). Bonato L. (007). Mone and Inflation in the Islaic Republic of Iran. [Online] Available: (Ma, 007) Bullard, J. (99). Tie-varing paraeters and non-convergence to rational expectations under least square learning. Econoic Letters. 0() Chen P. & Flaschel P. (006). Measuring the Interaction of Wage and Price Phillips Curves for the U.S. Econo. Studies in Nonlinear Dnaics & Econoetrics. 0() Evans, G. & Honkapohja S. (00). Learning and expectations in acroeconoics, Princeton Universit Press, Princeton, (Chapter ). Lahiri, K. (98). The econoetrics of inflationar expectations, North-Holland Pub. Co. (Chapter ). Liu O. & Adedeji O. (000). Deterinants of Inflation in the Islaic Republic of Iran: A Macroeconoic Analsis. [Online] Available: (Jul, 000). Orphanides A. and Willias J. C. (00). Iperfect knowledge, inflation expectations and onetar polic. [Online] Available: (Ma, 00). Sargent, T. J. (999). The Conquest of Aerican Inflation, Princeton Universit Press. (Chapter ). Turnovsk, S.J. & Wachter, M. L. (97). A Test of the "Expectations Hpothesis" Using Directl Observed Wage and Price Expectations. The Review of Econoics and Statistics, MIT Press. 5() Valadkhani, A. (006). Macroeconoetric Modelling in an Oil-Exporting Countr: The case of Iran. The Journal of Energ and Developent. () Table. Descriptive Statistics of the data ( ) Variable Mean Maxiu Miniu Std. dev. Wage growth arte (%) (W) Uneploent rate (%) (U) Output Gap(%) (OG) Inflation rate (%) ( ) Real GNP growth rate (%) ( ) Noinal growth rate of M (%) ( M ) Real growth rate of M (%) ( / Real growth rate of governent expenditure (%) ( g ) Real growth rate of consuption (%) ( c ) Iport price growth rate (%) ( ) Productivit growth rate (%) ( X ) Published b Sciedu Press 9 ISSN E-ISSN 97-60X

6 Business and Manageent Research Vol., No. ; 0 Table. Endogenous and Exogenous Variables under Scenario I (%) Variable Variable W M U M / P OG g e X.69.6 Y Table. Endogenous Variables under three Scenarios (%) Scenario I Scenario II Scenario III Variable Variable Variable w 9.5 w 0.7 W Y Published b Sciedu Press 9 ISSN E-ISSN 97-60X