1

THE ACADEMY OF ECONOMIC STUDIES BUCHARESTDOCTORAL SCHOOL OF FINANCE AND BANKING

DISSERTATION PAPERMoney Demand in Romania

Student: DUMITRU IONUŢSupervisor: PhD. Professor MOISĂ ALTĂR

BUCHAREST, JUNE 2002

2

The role of the demand for money in the transmission mechanism of monetary policy

Basic theoretical approaches to examination of the money demand

Money demand estimates in Eastern Europe

Money demand modelling in Romania

Conclusions

STRUCTURE:

3

1. Quantity theory

“Equation of exchange” –Irving Fischer’s equation

“Cambridge approach or cash balance approach”

2. Keynesian theory

3. Neo-Keynesian theory – Baumol and Tobin

4. Post-Keynesian theories

5. Modern monetarist approach - Milton Friedman

Basic theoretical approaches to examination of the money demand

4

MONEY DEMAND ESTIMATES IN EASTERN EUROPE

Klacek and Smidkova (1995) estimated the long-run demand for broad and narrow money in the Czech Republic

Van Aarle and Budina (1996) estimated money demand and specifically the effect of currency substitution using the portfolio balance approach for Poland, Hungary, Romania, and Bulgaria

Arlt, Guba, Radkovsky, Sojka and Stiller (2001) estimated money demand for Czech Republic in period 1994-2000.

Erwin Nijsse and Elmer Sterken (1996) estimated a household money demand function for Poland from 1969 to 1995.

Antoni Chawluk (2000) analysed the same household money demand and consumption for Poland.

5

MONEY DEMAND MODELLING IN ROMANIA

);;( xryfM d Typically, in the long run:y is a scale variable measuring the level of economic activity

r – a vector of opportunity cost variables

x is a vector of other variables (including dummy variables)

Money balances are measured as real Lei M2 (M2 less foreign currency deposits - in order to point out currency substitution). Why M2?– because NBR uses M2 as intermediary target.

As a scale variable - the real industrial production index (deflated by the consumer price index - CPI) as a proxy for GDP which is not calculated monthly in Romania.

Monthly data between January 1996 and March 2002

6

Deposit interest rate for non-bank clients – measures the rate of return on lei time deposits (proxy for return of M2).T-bills interest rate – measures the rate of return on assets outside of M2.Expected inflation rate – proxied by current inflation rate – captures the return on real assets and the substitution between monetary assets and real assets (goods).Expected depreciation of the lei-dollar exchange rate - captures the return on holding US dollars and currency substitution. The actual depreciation was used as proxy for the expected rate.

Opportunity costs:

7

Three specifications (models) are analysed

I. Money demand= f (output, deposit interest rate, T-bills interest rate, inflation) – closed economy model

II. Money demand= f (output, deposit interest rate, T-bills interest rate, inflation, depreciation of exchange rate) - open economy model

III. Money demand= f (output, deposit interest rate, T-bills interest rate, exchange rate) - open economy model

REMARK: As many of the series exhibit regular seasonal patterns, it is necessary to take account of the seasonal factors in the estimation. This is done in two ways:

1) using seasonal adjusted data through Tramo-Seats procedure (Eviews 4.0)

2) using raw series with monthly seasonal dummies (centred dummy)

8

Variable ADF test PP test

Real broad money* I(1) C I(1) C

Real output* I(1) C I(1) C

Exchange rate* I(1) C T I(1) C T

Exchange rate depreciation I(1) C or I(0) C I(0) C

Inflation I(1) C or I(0) C I(1) C or I(0) C

Deposit interest rate I(1) C T I(1) C T

T-bills interest rate I(1) C T I(1) C T

Seasonally adjusted variables

Real broad money* I(1) C I(1) C

Inflation I(1) C or I(0) C I(1) C or I(0) C

Real output* I(1) C I(1) C

UNIT ROOT TESTS

The cointegration technique is appropriate to estimate the long run money demand. Table suggests that none of the variables is integrated of order 2 (I(2)) or more. Inflation and depreciation of exchange rate is probably I(0). This does not mean that they should be excluded from cointegrating vector (Dickey and Rossana(1994), and Enders(2000)).

9

I have used multivariate Johansen procedure for cointegration.

The number of lags included in cointegration tests was determined by estimating a VAR with variables of interest and using criteria like LR, FPE, AIC, SC and HQ. If the optimum lag in VAR is p, then the VEC is estimated with p-1 lags.

For each cointegration test I have obtained one cointegrating vector at 1 or 5%.

Data through end-September 2001 were used for estimations, with the remaining observations reserved for out-of-sample forecasting (6 months).

REMARKS:

10

  Output Deposit interest rate T-bills interest rate Inflation

  Coef. SE t Coef. SE t Coef. SE t Coef. SE t

I 1.39 0.49 2.78 3.52 0.95 3.69 -2.13 0.55 -3.85 -0.48 0.18 -2.65

II 1.33 0.19 6.73 2.92 0.4 7.16 -0.65 0.19 -3.36 -0.31 0.07 -4.31

III 1.46 0.21 6.73 3.57 0.29 12.05 -1.27 0.15 -7.97 -0.47 0.05 -9.00

  Depreciation Exchange rate Speed of adjustment RMSE 4/

  Coef. SE t Coef. SE t Coef. SE t Stat. Dyn.

I             -0.04 0.01 -2.43 0.02 0.09

II -0.46 0.09 -4.98       -0.10 0.02 -4.17 0.03 0.12

III       -0.34 0.2 1.69 -0.11 0.02 -4.54 0.02 0.08

The results for seasonally adjusted data:

All the estimated coefficients are statistically significant at 1%.

Estimates for each of the models have the anticipated signs (according with economic theory) and there is relatively little variation in the estimates of key parameters.

As magnitude, the estimated coefficients are similar with those obtained in other empirical works performed in other countries from Central and Eastern Europe

11

Discussion on estimated coefficients The output elasticity in the long run relationship is greater than one as was expected for broad money aggregate. Output decrease in Romania, especially in the first half of the last decade, was accompanied by a demonetization.

Like magnitude, income elasticity does not differ significantly from 1:

)1(2 2.28[0.12]), which is consistent with the quantity theory of money.

The differences in size between interest rates coefficients in models I and II may be due to the fact that multi-collinearity problem becomes more acute in the case of model II due to the close correlation between exchange rate depreciation and inflation.

Deposit interest rate semi-elasticity of money demand is positive and has an opposite sign to the T-bills interest rate semi-elasticity.

The exchange rate depreciation coefficient has a negative sign (consistent with economic theory) which indicates the existence of a currency substitution in Romania.

12

The coefficient of depreciation is relatively small. This is explained by the fact that on average, the national currency appreciated in real terms excepting the 1997 and 1999 shocks, and the deposit interest rate was on average greater than the yield for foreign exchange deposits, which was very obvious since 2001.

The inflation semi-elasticity of money demand is negative, which involves a substitution between monetary assets and real assets. This is not surprising for a country like Romania in which financial assets outside broad money are limited and the agents hold significant quantities of real assets.

The coefficients representing speed of adjustment indicate relatively rapid adjustment of real money demand to disequilibria. The negative sign of speed of adjustment suggests that if in the previous month the money demand exceeded the long term level, in the current month money demand would decrease.

13

Weak-exogeneity tests

Weak exogeneity hypothesis is accepted both separately (for output, deposit interest rate, inflation, depreciation and exchange rate) and jointly. It is rejected for broad money and T-bills. The deposit interest rate weak exogeneity suggests that it is determined outside the system (it is not caused by money demand, but it determines money demand). The T-bills interest rate is not weak exogenous and it reacts to the money demand disequilibria from the long term level. The relationship between inflation and broad money is from inflation to broad money and not vice versa, the inflation being weak exogenous for the money demand. It is not the rise in broad money which generates inflation, but it is the broad money which is accommodated to inflation as this monetary aggregate rises to reach the equilibrium. Consequently, we can state that inflation is not a monetary phenomenon.

14

REMARK:The estimates using unadjusted data (with seasonally centred dummy – as Johansen suggested) were not consistent in economical and econometrical terms unless in the case of closed economy model, the other models leading to statistically insignificant coefficients.

The unrestricted cointegrating relation:

-2

-1

0

1

2

3

4

1996 1997 1998 1999 2000 2001

ECM1

-2

-1

0

1

2

3

4

1996 1997 1998 1999 2000 2001

ECM2

-2

-1

0

1

2

3

4

1996 1997 1998 1999 2000 2001

ECM3

-4

-2

0

2

4

6

1996 1997 1998 1999 2000 2001

ECM1dummy

The figures show that the deviation of money demand from its long term level for the three models with seasonally adjusted data and model I with unadjusted data is stationary, so we can use it in an error correction mechanism. This deviation is relatively small, excepting 1997 when in the first three months there was an obvious monetary overhang, and in the next two months a shortfall occurred due to pressures made by NBR to rise the interest rates, after which the situation came back to normal.

15

The stability of parameters in the long run equilibrium:

In figures are presented the CUSUM tests, recursive residuals, N-step forecast test and the recursive coefficients.

-.12

-.08

-.04

.00

.04

.08

.12

1999 2000 2001

Recursive Residuals ± 2 S.E.

-20

-10

0

10

20

1999 2000 2001

CUSUM 5% Significance

.00

.04

.08

.12

-.12

-.08

-.04

.00

.04

.08

.12

1999 2000 2001

N-Step Probability Recursive Residuals

I.

II.

III.

-.12

-.08

-.04

.00

.04

.08

.12

1998 1999 2000 2001

Recursive Residuals ± 2 S.E.

-20

-10

0

10

20

1998 1999 2000 2001

CUSUM 5% Significance

.00

.04

.08

.12

-.15

-.10

-.05

.00

.05

.10

.15

1998 1999 2000 2001

N-Step Probability Recursive Residuals

-.06

-.04

-.02

.00

.02

.04

.06

99:01 99:07 00:01 00:07 01:01 01:07

Recursive Residuals ± 2 S.E.

-20

-15

-10

-5

0

5

10

15

20

99:01 99:07 00:01 00:07 01:01 01:07

CUSUM 5% Significance

.00

.04

.08

.12

-.08

-.04

.00

.04

.08

99:01 99:07 00:01 00:07 01:01 01:07

N-Step Probability Recursive Residuals

16

-.12

-.08

-.04

.00

.04

.08

.12

99:01 99:07 00:01 00:07 01:01 01:07

ECM1_1 ± 2 S.E.

-1.0

-0.5

0.0

0.5

1.0

1.5

2.0

2.5

99:01 99:07 00:01 00:07 01:01 01:07

DLM2R_SA_1 ± 2 S.E.

-1.5

-1.0

-0.5

0.0

0.5

1.0

1.5

99:01 99:07 00:01 00:07 01:01 01:07

DLM2R_SA_2 ± 2 S.E.

-1.5

-1.0

-0.5

0.0

0.5

1.0

1.5

99:01 99:07 00:01 00:07 01:01 01:07

DLM2R_SA_3 ± 2 S.E.

-2.0

-1.6

-1.2

-0.8

-0.4

0.0

0.4

0.8

99:01 99:07 00:01 00:07 01:01 01:07

DLM2R_SA_4 ± 2 S.E.

-1.2

-0.8

-0.4

0.0

0.4

0.8

99:01 99:07 00:01 00:07 01:01 01:07

LYRIBF_SA_1 ± 2 S.E.

-1.2

-0.8

-0.4

0.0

0.4

0.8

99:01 99:07 00:01 00:07 01:01 01:07

LYRIBF_SA_2 ± 2 S.E.

-0.8

-0.4

0.0

0.4

0.8

1.2

1.6

99:01 99:07 00:01 00:07 01:01 01:07

LYRIBF_SA_3 ± 2 S.E.

-1.0

-0.8

-0.6

-0.4

-0.2

0.0

0.2

0.4

0.6

99:01 99:07 00:01 00:07 01:01 01:07

LYRIBF_SA_4 ± 2 S.E.

-1.5

-1.0

-0.5

0.0

0.5

1.0

1.5

99:01 99:07 00:01 00:07 01:01 01:07

DDP_1 ± 2 S.E.

-.8

-.6

-.4

-.2

.0

.2

.4

.6

.8

99:01 99:07 00:01 00:07 01:01 01:07

DDP_3 ± 2 S.E.

-.6

-.4

-.2

.0

.2

.4

99:01 99:07 00:01 00:07 01:01 01:07

DDP_4 ± 2 S.E.

-.3

-.2

-.1

.0

.1

.2

99:01 99:07 00:01 00:07 01:01 01:07

DDTS_1 ± 2 S.E.

-.3

-.2

-.1

.0

.1

.2

99:01 99:07 00:01 00:07 01:01 01:07

DDTS_2 ± 2 S.E.

-.3

-.2

-.1

.0

.1

.2

.3

99:01 99:07 00:01 00:07 01:01 01:07

DDTS_3 ± 2 S.E.

-.2

-.1

.0

.1

.2

.3

.4

99:01 99:07 00:01 00:07 01:01 01:07

DDTS_4 ± 2 S.E.

-.08

-.04

.00

.04

.08

.12

99:01 99:07 00:01 00:07 01:01 01:07

DP_SA_1 ± 2 S.E.

-.06

-.04

-.02

.00

.02

.04

.06

.08

.10

99:01 99:07 00:01 00:07 01:01 01:07

DP_SA_2 ± 2 S.E.

-.04

-.02

.00

.02

.04

.06

.08

99:01 99:07 00:01 00:07 01:01 01:07

DP_SA_3 ± 2 S.E.

-.04

-.02

.00

.02

.04

.06

.08

99:01 99:07 00:01 00:07 01:01 01:07

DP_SA_4 ± 2 S.E.

Recursive coefficients of the short run unrestricted ECM (model I):

According to the tests performed, the estimated coefficients are constant over time, although in 1997 there are some signs of instability.

Because of the insufficient number of observations until 1997:01 (12 observations), the models cannot be estimated separately for the two sub-periods; this is the reason why a Chow test for a structural break in 1997 cannot be performed.

17

In order to test if the 1997 shock produced only a one time jump in the determinants of money demand and left them unmodified, we re-estimated the cointegrating relation for the 1997:06-2001:09 period. For the long run we have obtained parameters similar to those obtained for the entire sample. For example, in model II, if we restrict all the long run coefficients estimated for the 1997:06-2001:09 sample to be the same with those estimated for the entire sample we cannot reject the null hypothesis:

REMARK:

)5(2 5.73[0.34].

18

Short run error correction model (ECM)

o The estimated cointegrating equations include the factors affecting the real money demand in the long run. In the short run deviations from these relations can occur.

I estimate a short run parsimonious error correction model as follows:

ttitjj i

ji ECMVCRLM 11,

5

1

6

1

2

19

Dependent Variable: D(LM2R_SA)

Method: Least Squares

Sample(adjusted): 1996:03 2001:09

Included observations: 67 after adjusting endpoints

Variable Coefficient

Std. Error t-Statistic Prob.

ECM1_1 -0.013942 0.004046 -3.446164 0.0010

DLM2R_SA_1 0.313701 0.115123 2.724918 0.0083

DDTS -0.026417 0.011747 -2.248838 0.0281

DP_SA -0.014734 0.001866 -7.894747 0.0000

C -0.005467 0.002521 -2.168777 0.0339

R-squared 0.667695 Mean dependent var -0.007540

Adjusted R-squared 0.646256 S.D. dependent var 0.032469

S.E. of regression 0.019311 Akaike info criterion -4.984542

Sum squared resid 0.023122 Schwarz criterion -4.820013

Log likelihood 171.9822 F-statistic 31.14393

Durbin-Watson stat 2.144583 Prob(F-statistic) 0.000000

Using a general-to-specific methodology (D. Hendry) by eliminating insignificant lags we obtain a parsimonious model:

Like in the unrestricted model, in the restricted model the error correction term has a negative sign and is statistically significant. In other words, if there is an excess of money in the current month, in the next month the agents will reduce their money holdings.

20

While the changes in output and in deposit interest rate do not affect the short run changes in money demand, the money demand is influenced in the short run by changes in T-bills interest rate and inflation.

REMARK:

-.08

-.06

-.04

-.02

.00

.02

.04

.06

.08

1997 1998 1999 2000 2001

Recursive Residuals ± 2 S.E.

-30

-20

-10

0

10

20

30

1997 1998 1999 2000 2001

CUSUM 5% Significance

.00

.04

.08

.12

-.08

-.04

.00

.04

.08

1997 1998 1999 2000 2001

N-Step Probability Recursive Residuals

-.05

-.04

-.03

-.02

-.01

.00

.01

.02

1997 1998 1999 2000 2001

ECM1_1 ± 2 S.E.

-0.8

-0.4

0.0

0.4

0.8

1.2

1997 1998 1999 2000 2001

DLM2R_SA_1 ± 2 S.E.

-.08

-.04

.00

.04

.08

1997 1998 1999 2000 2001

DDTS ± 2 S.E.

-.05

-.04

-.03

-.02

-.01

.00

.01

1997 1998 1999 2000 2001

DP_SA ± 2 S.E.

-.03

-.02

-.01

.00

.01

.02

1997 1998 1999 2000 2001

C ± 2 S.E.

The coefficients from the parsimonious model are more stable, the error bands decreasing rapidly. After 1997, the coefficients from the parsimonious ECM are virtually constant.

In figures-the CUSUM tests, recursive residuals, N-step forecast test and the recursive coefficients.

21

Comparing the unrestricted model to parsimonious model, we observe that the last is better in qualitative terms, and this is confirmed by the normality tests of residuals.

0

1

2

3

4

5

6

7

8

9

-0.04 -0.02 0.00 0.02 0.04

Series: ResidualsSample 1996:03 2001:09Observations 67

Mean -1.42E-18Median -0.001275Maximum 0.046782Minimum -0.043691Std. Dev. 0.018717Skewness 0.179114Kurtosis 2.529239

Jarque-Bera 0.976923Probability 0.613570

0

1

2

3

4

5

6

7

8

-0.06 -0.04 -0.02 0.00 0.02 0.04

Series: ResidualsSample 1996:06 2001:09Observations 64

Mean -2.49E-18Median 0.000910Maximum 0.036619Minimum -0.060841Std. Dev. 0.018561Skewness -0.386101Kurtosis 3.321162

Jarque-Bera 1.865176Probability 0.393534

Normality test of residual for unrestricted ECM – model I

Normality test of residual for parsimonious ECM – model I

22

Forecasting ability

• Extending the model for the period October 2001-March 2002 we can make a series of judgments regarding the stance of monetary policy during this period, although this is subject to errors.

M2 actual vs. fitted model ECM parsimonious

8.9

9.0

9.1

9.2

9.3

9.4

9.5

9.6

9.7

1996 1997 1998 1999 2000 2001

actual fitted

9.04

9.06

9.08

9.10

9.12

9.14

9.16

9.18

2001:10 2002:01

actual dinamic static8.9

9.0

9.1

9.2

9.3

9.4

9.5

9.6

9.7

1996 1997 1998 1999 2000 2001

fitted

Forecast: LM2R_SAFActual: LM2R_SAForecast sample: 1996:01 2002:03Adjusted sample: 1996:03 2002:03Included observations: 73

Root Mean Squared Error 0.018253Mean Absolute Error 0.014995Mean Abs. Percent Error 0.162360Theil Inequality Coefficient 0.000991 Bias Proportion 0.002481 Variance Proportion 0.002368 Covariance Proportion 0.995151

forecasts suggest that the models provide a reasonable approximation of the actual money demand process during 1996-2002

the forecast for 2001:10-2002:03 can give us a clue about the stance of monetary policy, indicating that on average the actual real money demand was larger than the estimated money demand during this period, which generated a monetary overhang, the monetary policy being more relaxed.

23

A state space model for the money demand in Romania

I estimated a state space model (time-varying parameters) model using Kalman filters on the demand for money in Romania, following an approach similar to Bomhoff (1991) and Stracca (2001).

a time-varying parameters model can underline possible changes in parameters over time.

it can asses the impact of financial innovation onto money holdings.

Advantages:

tttttt

d

t pDTSDPykm ****

ttt u 110

Concretely, I specified a state space model where I explicitly allowed the deposit interest rate elasticity of money demand to vary. The model is specified as:

24

The freely estimate for the output elasticity tends to give a value close to 1.35. This is also close to the value of 1.39 estimated by Johansen procedure.

The time-varying-parameters model is estimated by means of a Kalman filter over the full sample period. This procedure requires maximizing the likelihood function using an optimization algorithm (in particular the BHHH – Berndt-Hall-Hall-Hausman algorithm).

The model appears to be well specified and the residual appears to be stationary. Being the model specified in terms of the long run relationship, the residual have some positive autocorrelation (Q statistic Q(6)=14.35[0.00]), reflecting the existence of adjustments costs in bringing monetary holdings to the desired equilibrium value.

REMARKS:

-4

-3

-2

-1

0

1

2

3

4

1996 1997 1998 1999 2000 2001

residual

Residual from Kalman filter

25

The deposit rate elasticity of the money demand

-20

-10

0

10

20

30

40

50

60

70

1996 1997 1998 1999 2000 2001

DP ± 2 RMSE

Filtered State DP Estimate

A reduced but noticeable increase in absolute value of the deposit interest rate elasticity of money demand is visible beginning from the first half of 2000. This means that in the context of a progress towards a more reduced and predictable inflation, reduced and stable interest rates, the preference for liquidity of the agents increases. Another remark is related to the sharp decline in the deposit interest rate elasticity of money demand at the beginning of 1997. This is due to the high level of deposit interest rate (up to120-130% for term deposits) for this period which caused the money demand not to vary significantly at one percent change in the deposit interest rate.

26

CONCLUSIONS Cointegration analysis reveals that real broad money balances are determined in the long run by real output, deposit interest rate, T-bills interest rate, inflation and exchange rate depreciation.In Romania, as we have seen, the inflation is weekly exogenous for money demand, which means that inflation is not a monetary phenomenon. The opportunity cost variables carry the expected sign according to economic theory. The depreciation of exchange rate elasticity of the money demand indicates currency substitution in Romania.In the short run, the money demand is not influenced by changes in output and deposit interest rate, but changes in T-bills interest rate and inflation. The demand for real money is stable for the period analyzed. Further evidence in support of the stability of the model derives from its positive forecasting performance. The important findings are that both in the long and in the short-run the demand for real M2 is stable providing justification for the monetary authorities to use M2 aggregate like intermediary target in conducting his policy.

27

1. Arestis Ph., (1988), “The Demand for Money in Small Developing Economies: An Application of the Error Correction Mechanism” In: Ph. Arestis (ed.): ”Contemporary Issues in Money and Banking. Chaltenham, Edward Elgar

2. Arlt, J., M. Guba, S. Radkovsky, M. Sojka, V. Stiller, (2001), “Influence of selected factors on the the demand for money 1994-2000”, Czech National Bank Working Paper No. 30, Praha

3. Ball, L. (2001), “Another Look at Long-Run Money Demand”, Journal of Monetary Economics, 47, 1, pp. 31-44.

4. Bomhoff, E. (1991), “Stability and Velocity in the Major Industrial Countries: A Kalman Filter Approach”, IMF Staff Papers, 38, 3, pp. 626-642.

5. Brand, C. and N. Cassola (2000), “A Money Demand System for Euro Area M3”, ECB Working Paper Series n. 39.

6. Budina, N. and S. van Wijnbergen (1998), "Fiscal Deficits, Monetary Reform and Inflation Stabilization in Romania", World Bank, Development Research Group

7. Chadha, J. S., Haldane, A. G. and N. G. J. Janssen (1998), “Shoe-Leather Costs Reconsidered”, Economic Journal, 108, 447, pp. 363-382.

8. Charemza, W., D., "New Directions in Econometric Practice: General to Specific Modeling, Cointegration and Vector Autoregression", E. Elgar Publishing Ltd. 1992.

9. Coenen, G. and J. L. Vega (1999),”The Demand for M3 in the Euro Area”, ECB Working Paper n. 6.

10. Cuthbertson, K., D. Bredin, (2001), “Money demand in Czech Republic since transition”, Journal of Policy Reform

REFERENCES

28

11.Dickey, D. A. and Rossana, R. J. (1994), “Cointegrated time series: A guide to estimation and hypothesis testing”, Oxford Bulletin of Economic and Statistics, 56 (3), 325-53.

12.Egoume-Bosogo, P., (2000), “Money demand in Guyana”, IMF Working Paper No. 11913.Enders, W., (2000), “Applied econometric time series”, Iowa State University, John

Wiley & Sons, Inc14.Engle R. F., Granger C. W. J. (1987) – “Cointegration and Error Correction:

Representation, Estimation and Testing”, Econometrica, 5515.Ericsson, N. R. (1998), “Empirical Modeling of Money Demand”, Empirical Economics,

23, 3, pp. 295-315.16.Ericsson, N. R., D. F. Hendry, K. M. Prestwich, (1998), “Friedman and Schwartz (1982)

resisted: Assessing annual and phase-average models of money demand in the United Kingdom”, Empirical Economics 23: 401-415

17.Ericsson, N.R. (1999), “Empirical modeling of money demand”, in H. Lütkepohl and J. Wolters (eds.), Money Demand in Europe, (Physica-Verlag) Heidelberg, 29-49.

18.Ericsson, N.R., D.F Hendry and G.E. Mizon (1998), “Exogeneity, cointegration and economic policy analysis”, Journal of Business and Economic Statistics, 16, 370-387.

19.Favero, A. C., (2001), “Applied macro econometrics”, Oxford University Press20.Greene, W. (1993), “Econometric Analysis”, second edition, Macmillan Publishing

Company, New York, N.Y.21.Hamilton, D. J., (1994), “Time series analysis”, Princeton University Press22.Hendry, F. D., and K. Juselius (2000), “Explaining Cointegration Analysis: Part II”,

Energy Journal 21

29

23.Johansen S., Juselius K., (1990), “Maximum Likelihood Estimation and Inference on Cointegration - with Applications to Simultaneous Equations and Cointegration”, Journal of Econometrics, 69

24.Judd, J.P., and J.L. Scadding, (1982), “The Search for a Stable Money Demand Function: A survey of the post-1973 literature,” Journal of Economic Literature, Vol. 20, No. 3, pp. 993-1023

25.Jusoh, M. (1987), „Inflationary Expectations and the Demand for Money in moderate inflation: Malayesian evidence”, Journal Ekonomi Malayesia, Vol. 15, pp. 3-14.

26.Lucas, R. (2000), “Inflation and Welfare”, Econometrica, 68, 2, pp. 247-274.27.Lütkepohl, H. and H.-E. Reimers (1992), “Impulse response analysis of cointegrated

systems”, Journal of Economics Dynamics and Control, 16, 53-78.28.Lütkepohl, H., Teräsvirta, T. and J. Wolters (1999) – “Investigating Stability and

Linearity of a German M1 Demand Function”, Journal of Applied Econometrics, 14, pp. 511-525.

29.Mullighan, C. and X. Sala-i-Martin (1996), "Adoption of Financial Technologies: Technologies and Implications for Money Demand and Monetary Policy", NBER Working Paper n. 5504.

30.Özmen, E (1998), "Is Currency Seignorage Exogenous for Inflation Tax in Turkey?", Applied Economics, 30, 545 – 552

31.Patterson, K., (2000), “An introduction to applied econometrics: a time series approach”, Palgrave

32.Petursson, T., (2001), “The representative household’s demand for money in a cointegrated VAR model”, Central Bank of Iceland, Working Paper No. 12

30

33.Sidrauski, M. (1967), "Rational Choice and Patterns of Growth in a Monetary Economy", American Economic Review, 57, pp. 534-544.

34.Simmons, R. (1992), “An error-correction approach to demand for money in five African developing countries”, Journal of Economic Studies, Vol 19, pp. 29-47

35.Soto, R., M. Tapia, (2001), “Seasonal cointegration and the stability of the demand for money”, Central Bank of Chile, Working Paper No. 103

36.Sriram S. S. (1999a), “Survey of Literature on Demand for Money: Theoretical and Empirical Work with Special Reference to Error-Correction Models”, IMF Working Paper No. 64, Washington

37.Sriram S. S. (1999b), “Demand for M2 in an Emerging-Market Economy: An error-correction Model for Malaysia”, IMF Working Paper No. 173, Washington

38.Stracca, L., (2001), “The functional form of the demand for euro area M1”, ECB Working Paper 51

39.Van Aarle, B. and Budina, N., (1996), “Currency Substitution and Seignorage in Eastern Europe”, The Journal of Policy Reform, Vol. 1, pp. 279-98.

40.Walsh, C. E., (1998), "Monetary Policy and Theory", MIT Press, Cambridge41.*** "Romania: Selected Issues and Statistical Appendix", IMF, 200142.*** International Monetary Fund, (2001) „IMF Country Report No. 01/204 - Romania:

Request for a Stand-By Arrangement”, November43.*** International Monetary Fund, (2001) „IMF Country Report No. 16/01 - Romania:

Selected Issues and Statistical Appendix”, January 44.*** National Bank of Romania, “Monthly bulletin” from period 1996-2002, Bucharest,

Romania45.*** National Bank of Romania, “Annual Report” from period 1997-2002, Bucharest,

Romania