Abstract
This Selected Issues paper analyzes empirically the main determinants of Hungary’s inflation rate during 1990–96. Although there exist a number of possible methodologies to analyze this issue, the one proposed in the paper takes explicit account of the time-series properties of the variables that are potential candidates for explaining Hungary’s inflation performance. This leads to the specification of a long-term equation, linking consumer prices to a number of macroeconomic variables as well as to proxies for relative price shocks. The paper also examines the external current account and net foreign assets in Hungary.
VI. Capital Flows, Reverse Currency Substitution, and Sterilized Intervention 1/
1. Introduction
Following the introduction of the March 1995 austerity package and the adoption of a preannounced crawling peg exchange regime (see EBS/96/18, 2/5/1996, for details), the conduct of monetary policy has been complicated by strong capital inflows and, more importantly, by the conversion of foreign exchange assets into forint by both residents and nonresidents. In order to maintain the forint within its +/-2¼ percent band and yet keep interest rates at a sufficiently high level in view of the persisting inflation, the NBH has engaged in sterilization operations. This chapter examines some of the factors behind these recent trends and assesses the effectiveness of sterilization operations using an empirical framework.
2. Short-term capital inflows and reverse currency substitution
a. Interest rate differentials
Interest rate differentials have played a key role in inducing currency substitution and capital flows. High interest rate differentials in favor of forint-denominated assets persisted after the introduction of the March 1995 policy package. The interest rate premium, after adjusting for the expected movements in the exchange rate, remained in the 10-15 percentage points range until December 1995, before gradually declining to some 5 percentage points by May 1996 (Chart 22). Initially, part of this premium may have reflected the risks associated with the short track record of the Central Bank in maintaining a preannounced exchange rate regime within a relatively narrow band. It may have also reflected the segmentation of the Hungarian financial markets in which nonresidents were not able to fully exploit the arbitrage opportunities between assets denominated in foreign and domestic currency. Of particular importance was the fact that non-residents were not allowed to buy government instruments that had a remaining maturity of less than 12 months. (This restriction was partially relaxed in early 1996, see Chapter VII). 2/
HUNGARY: INTEREST RATE PREMIUM 1/
(In percent)
Citation: IMF Staff Country Reports 1996, 109; 10.5089/9781451817829.002.A006
Sources: National Bank of Hungary; and staff calculations.1/ 12-month T-bill relative to German interbank rate, assuming a continuation of the 1.2 percent monthly rate of crawl for the forint after July 1996.HUNGARY: INTEREST RATE PREMIUM 1/
(In percent)
Citation: IMF Staff Country Reports 1996, 109; 10.5089/9781451817829.002.A006
Sources: National Bank of Hungary; and staff calculations.1/ 12-month T-bill relative to German interbank rate, assuming a continuation of the 1.2 percent monthly rate of crawl for the forint after July 1996.HUNGARY: INTEREST RATE PREMIUM 1/
(In percent)
Citation: IMF Staff Country Reports 1996, 109; 10.5089/9781451817829.002.A006
Sources: National Bank of Hungary; and staff calculations.1/ 12-month T-bill relative to German interbank rate, assuming a continuation of the 1.2 percent monthly rate of crawl for the forint after July 1996.The sizable interest rate premium on domestic currency denominated assets that prevailed in Hungary during most of 1995 and early 1996 was a very strong incentive for banks, firms, and individuals to shift their asset holdings away from foreign currency and to borrow abroad. This resulted in the conversion of foreign currency deposits to domestic assets, and short term inflows through the balance of payments.
b. Short-term capital inflows
The short term capital inflows observed particularly in late 1995 and early 1996 (Table 16) mainly reflected the net borrowing operations of banks and enterprises that were either exploiting arbitrage opportunities offered by the high interest rate premium on forint denominated assets or using, whenever available, a cheaper source of funds. Errors and omissions in the balance of payments were also large, possibly reflecting disguised forms of short term capital inflows.
HUNGARY: Short-term Capital, Foreign Currency Deposits by Commercial Banks at the NBH, Holdings of Government Bonds, and Reverse Repo Operations, December 1994 - May 1996
(In millions of U.S. dollars)
HUNGARY: Short-term Capital, Foreign Currency Deposits by Commercial Banks at the NBH, Holdings of Government Bonds, and Reverse Repo Operations, December 1994 - May 1996
(In millions of U.S. dollars)
| Short-term Capital (net) | Changes in Foreign Currency Deposits | (A)-(B)-(C) | Stock of Foreign Currency Deposits | Changes in the Holdings of Government Bonds by NBH | Reverse Repo Operations (flow) | -(D)+(E) | ||||
|---|---|---|---|---|---|---|---|---|---|---|
| Households | Banks | |||||||||
| (A) | (B) | (C) | (D) | (E) | ||||||
| December 1994 | -187 | … | -205 | 4658 | 84 | 6 | -78 | |||
| January 1995 | 217 | 191 | 483 | -457 | 5141 | -150 | -42 | 108 | ||
| February 1995 | -37 | 166 | 695 | -897 | 5836 | 143 | -5 | -148 | ||
| March 1995 | 388 | 125 | -437 | 700 | 5399 | 234 | 114 | -120 | ||
| April 1995 | -90 | -178 | -433 | 520 | 4967 | -204 | -114 | 90 | ||
| May 1995 | -89 | 80 | -164 | -6 | 4803 | -191 | 0 | 191 | ||
| June 1995 | 69 | 70 | -4 | 3 | 4798 | 59 | 1 | -58 | ||
| July 1995 | -66 | 7 | -21 | -52 | 4777 | -87 | 124 | 210 | ||
| August 1995 | 115 | 59 | -185 | 241 | 4593 | -156 | 160 | 316 | ||
| September 1995 | -96 | -59 | -353 | 316 | 4240 | -424 | -138 | 286 | ||
| October 1995 | -75 | 90 | -34 | -131 | 4205 | -183 | -72 | 111 | ||
| November 1995 | 500 | -8 | -58 | 567 | 4147 | -110 | 146 | 256 | ||
| December 1995 | 575 | 11 | 763 | -199 | 4910 | -156 | 177 | 333 | ||
| January 1996 | 200 | 14 | -339 | 525 | 4571 | -152 | 84 | 236 | ||
| February 1996 | 266 | -94 | -152 | 512 | 4419 | -158 | 370 | 529 | ||
| March 1996 | -21 | 38 | -292 | 233 | 4127 | -72 | 490 | 562 | ||
| April 96 | -309 | -68 | -389 | 148 | 3738 | 1 | 238 | 237 | ||
| May 1996 | 66 | 0 | -380 | 446 | 3358 | -75 | 158 | 233 | ||
| Memorandum Item: | ||||||||||
| Cumulative | ||||||||||
| March 1995 - December 1995 | 1231 | 198 | -926 | 1959 | -1218 | 398 | 1616 | |||
| January 1996 - May 1996 | 202 | -109 | -1552 | 1863 | -456 | 1340 | 1796 | |||
HUNGARY: Short-term Capital, Foreign Currency Deposits by Commercial Banks at the NBH, Holdings of Government Bonds, and Reverse Repo Operations, December 1994 - May 1996
(In millions of U.S. dollars)
| Short-term Capital (net) | Changes in Foreign Currency Deposits | (A)-(B)-(C) | Stock of Foreign Currency Deposits | Changes in the Holdings of Government Bonds by NBH | Reverse Repo Operations (flow) | -(D)+(E) | ||||
|---|---|---|---|---|---|---|---|---|---|---|
| Households | Banks | |||||||||
| (A) | (B) | (C) | (D) | (E) | ||||||
| December 1994 | -187 | … | -205 | 4658 | 84 | 6 | -78 | |||
| January 1995 | 217 | 191 | 483 | -457 | 5141 | -150 | -42 | 108 | ||
| February 1995 | -37 | 166 | 695 | -897 | 5836 | 143 | -5 | -148 | ||
| March 1995 | 388 | 125 | -437 | 700 | 5399 | 234 | 114 | -120 | ||
| April 1995 | -90 | -178 | -433 | 520 | 4967 | -204 | -114 | 90 | ||
| May 1995 | -89 | 80 | -164 | -6 | 4803 | -191 | 0 | 191 | ||
| June 1995 | 69 | 70 | -4 | 3 | 4798 | 59 | 1 | -58 | ||
| July 1995 | -66 | 7 | -21 | -52 | 4777 | -87 | 124 | 210 | ||
| August 1995 | 115 | 59 | -185 | 241 | 4593 | -156 | 160 | 316 | ||
| September 1995 | -96 | -59 | -353 | 316 | 4240 | -424 | -138 | 286 | ||
| October 1995 | -75 | 90 | -34 | -131 | 4205 | -183 | -72 | 111 | ||
| November 1995 | 500 | -8 | -58 | 567 | 4147 | -110 | 146 | 256 | ||
| December 1995 | 575 | 11 | 763 | -199 | 4910 | -156 | 177 | 333 | ||
| January 1996 | 200 | 14 | -339 | 525 | 4571 | -152 | 84 | 236 | ||
| February 1996 | 266 | -94 | -152 | 512 | 4419 | -158 | 370 | 529 | ||
| March 1996 | -21 | 38 | -292 | 233 | 4127 | -72 | 490 | 562 | ||
| April 96 | -309 | -68 | -389 | 148 | 3738 | 1 | 238 | 237 | ||
| May 1996 | 66 | 0 | -380 | 446 | 3358 | -75 | 158 | 233 | ||
| Memorandum Item: | ||||||||||
| Cumulative | ||||||||||
| March 1995 - December 1995 | 1231 | 198 | -926 | 1959 | -1218 | 398 | 1616 | |||
| January 1996 - May 1996 | 202 | -109 | -1552 | 1863 | -456 | 1340 | 1796 | |||
c. Reverse currency substitution
There has been a significant decline in the foreign currency deposits of banks at the NBH since March 1995, which became particularly marked from January 1996. Part of this decline is unrelated to reverse currency substitution but reflects regulatory changes on reserve requirements which were phased in between February and June 1996. 1/ However, a large part of the decline reflects the conversion of these deposits into domestic currency because of the higher rate of return on forint-denominated assets. The resulting open foreign exchange position of banks was offset mostly by an extensive use of foreign exchange futures contracts: as a result, weekly turnover in the futures market increased from an average of about $40 million in December 1995 to averages of $100-150 million in March-April 1996.
d. Sterilization
The conversion of foreign currency deposits into forints and short term capital inflows put substantial upward pressure on the forint. In fact, during late 1995 and early 1996 the Central Bank had to intervene heavily in the foreign exchange market in order to maintain the forint within the preannounced band (Chart 23). These operations were partly sterilized by the NBH. However, this did not prevent an acceleration in broad money, and a related decline in T-bill interest rates, and hence a drop in the interest premium (Chart 22). As a result, beginning in April 1996, the interest rate differential had shrunk to about 5 percentage points; this did not offer major opportunities for arbitrage between foreign currencies and the forint taking into account the risks related to the fluctuation band (+/-2¼ percent around the central parity).
HUNGARY: MARKET EXCHANGE RATE AND EXCHANGE RATE BANDS
(Forint per U.S. dollar)
Citation: IMF Staff Country Reports 1996, 109; 10.5089/9781451817829.002.A006
Sources: National Bank of Hungary; and Fund staff calculations.HUNGARY: MARKET EXCHANGE RATE AND EXCHANGE RATE BANDS
(Forint per U.S. dollar)
Citation: IMF Staff Country Reports 1996, 109; 10.5089/9781451817829.002.A006
Sources: National Bank of Hungary; and Fund staff calculations.HUNGARY: MARKET EXCHANGE RATE AND EXCHANGE RATE BANDS
(Forint per U.S. dollar)
Citation: IMF Staff Country Reports 1996, 109; 10.5089/9781451817829.002.A006
Sources: National Bank of Hungary; and Fund staff calculations.During the period January-May 1996 the changes in foreign currency deposits at the Central Bank, and the short term capital inflows recorded in the balance of payments were sterilized through the use of two main instruments: (a) reverse repo operations with commercial banks (over US$1.3 billion); (b) sale of government bonds (about $500 million) (Table 16 and Chart 24). 1/ Most recently, as the exchange rate-adjusted interest rate premium has declined to some 5 percent, the incentives for shifts into domestic assets have been reduced, with the practical elimination of the concomitant need for foreign exchange intervention by the Central Bank.
HUNGARY: SHORT TERM INFLOWS, FOREIGN CURRENCY DEPOSIT WITHDRAWAL, AND STERILIZATION OPERATIONS
Citation: IMF Staff Country Reports 1996, 109; 10.5089/9781451817829.002.A006
Source: Fund staff calculations.HUNGARY: SHORT TERM INFLOWS, FOREIGN CURRENCY DEPOSIT WITHDRAWAL, AND STERILIZATION OPERATIONS
Citation: IMF Staff Country Reports 1996, 109; 10.5089/9781451817829.002.A006
Source: Fund staff calculations.HUNGARY: SHORT TERM INFLOWS, FOREIGN CURRENCY DEPOSIT WITHDRAWAL, AND STERILIZATION OPERATIONS
Citation: IMF Staff Country Reports 1996, 109; 10.5089/9781451817829.002.A006
Source: Fund staff calculations.3. Empirical evidence
If domestic and foreign assets were perfect substitutes and there were no effective impediments to capital flows, then sterilized intervention would have little effect on domestic interest rates or the money supply. At best sterilization could buy time. However, in the case of Hungary, as noted above (also see Chapter VII), capital account restrictions were in place during 1995, and to a lesser extent, also in 1996. Given this, and the existence of the preannounced exchange rate peg, the empirical question is: how effective has the NBH sterilization been?
To answer this question, we use an empirical framework pioneered by Kouri and Porter (1974) and Argy and Kouri (1974). This methodology involves estimating an “offset” coefficient which measures the degree to which changes in net foreign assets (NFA) position offset the effect on money of changes in net domestic assets (NDA). This is a standard method of measuring the effectiveness of sterilization operations when the exchange rate is predetermined (see, for example. Schadler et al, 1993). The reduced form equation which is usually estimated is based on a structural model which includes money demand and money supply functions. More specifically, the offset coefficient is calculated by estimating the following reduced form:
where Δ denotes the change in a variable, both NFA and NDA are defined at the central bank level, Y is nominal output 2/, and i is the foreign interest rate. 3/ According to this model, if capital is mobile or domestic and foreign assets can be easily substituted, then a reduction in the NDA of the central bank, say, through selling of bonds, will lead to an offsetting increase in NFA holdings. When a1 is close to -1, then the central bank’s domestic money operations are fully undone and sterilization intervention would be inconsequential.
The above equation was estimated using monthly data for the period December 1993-May 1996. In order to capture the potential impact of the adoption of the crawling peg exchange regime on the parameters of the above model, two dummy variables were also added for the period March 1995-May 1996. The first, D, is a simple additive dummy taking the value of 1 from March 1995 onwards; and the second is a multiplicative dummy, D*ΔNDA, which captures any change in a1 after March 1995. The estimated parameters are presented in Table 17. Before discussing these estimates, a note of caution is warranted. Given the short sample period and the unavailability or limitations of data for some key variables, e.g. on nominal monthly output, the results cannot be taken as conclusive evidence of the effectiveness of sterilization.
HUNGARY: Regression Estimates of Offset Coefficients with ΔNFA as Dependant Variable 1/
(1993: 12 - 1996: 5)
All regressions are estimated using OLS. The standard errors are in parentheses and the variables which are significant at 5 percent level are marked with *.
Constant term.
Lagrange multiplier test of serial correlation, the critical value is 21.03.
HUNGARY: Regression Estimates of Offset Coefficients with ΔNFA as Dependant Variable 1/
(1993: 12 - 1996: 5)
| Independent Variables | |||||||||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| C 2/ | ΔNFA(-1) | ΔNDA | ΔNDA(-1) | D*ΔNDA | D | ΔY | ΔY(-1) | Δi | Δi(-1) | R-2 | Durbin Watson | LM(12) Test 3/ | Offset Before March 1995 | Offset After March 1995 | |
| I | -2.63 | -0.13 | -0.60 | … | -0.61 | 28.07 | … | … | … | … | 0.94 | 2.08 | 17.9 | -0.53 | -1.07 |
| (10.87) | (0.05)* | (0.16)* | (0.17)* | (13.23) | |||||||||||
| II. | -12.10 | -0.08 | -0.42 | … | -0.77 | 31.93 | … | 1.16 | … | … | 0.95 | 2.31 | 20.3 | -0.38 | -1.09 |
| (10.33) | (0.05) | (0.15)* | (0.17)* | (11.85)* | (0.45)* | ||||||||||
| III. | -10.65 | -0.24 | -0.38 | -0.19 | -0.79 | 34.01 | 0.01 | 1.26 | -31.14 | 24.99 | 0.95 | 1.82 | 20.7 | -0.46 | -1.09 |
| (12.08) | (0.17) | (0.17)* | (0.19) | (0.18)* | (13.05) | (0.57) | (0.53)* | (35.60) | (35.53) | ||||||
All regressions are estimated using OLS. The standard errors are in parentheses and the variables which are significant at 5 percent level are marked with *.
Constant term.
Lagrange multiplier test of serial correlation, the critical value is 21.03.
HUNGARY: Regression Estimates of Offset Coefficients with ΔNFA as Dependant Variable 1/
(1993: 12 - 1996: 5)
| Independent Variables | |||||||||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| C 2/ | ΔNFA(-1) | ΔNDA | ΔNDA(-1) | D*ΔNDA | D | ΔY | ΔY(-1) | Δi | Δi(-1) | R-2 | Durbin Watson | LM(12) Test 3/ | Offset Before March 1995 | Offset After March 1995 | |
| I | -2.63 | -0.13 | -0.60 | … | -0.61 | 28.07 | … | … | … | … | 0.94 | 2.08 | 17.9 | -0.53 | -1.07 |
| (10.87) | (0.05)* | (0.16)* | (0.17)* | (13.23) | |||||||||||
| II. | -12.10 | -0.08 | -0.42 | … | -0.77 | 31.93 | … | 1.16 | … | … | 0.95 | 2.31 | 20.3 | -0.38 | -1.09 |
| (10.33) | (0.05) | (0.15)* | (0.17)* | (11.85)* | (0.45)* | ||||||||||
| III. | -10.65 | -0.24 | -0.38 | -0.19 | -0.79 | 34.01 | 0.01 | 1.26 | -31.14 | 24.99 | 0.95 | 1.82 | 20.7 | -0.46 | -1.09 |
| (12.08) | (0.17) | (0.17)* | (0.19) | (0.18)* | (13.05) | (0.57) | (0.53)* | (35.60) | (35.53) | ||||||
All regressions are estimated using OLS. The standard errors are in parentheses and the variables which are significant at 5 percent level are marked with *.
Constant term.
Lagrange multiplier test of serial correlation, the critical value is 21.03.
Three alternative specifications, estimated by ordinary least squares (OLS) are reported in Table 17. 1/ The contemporaneous change in NDA is in all cases a significant variable in explaining the movements in the central bank’s NFA. 2/ The first equation simply includes ΔNDA, D, D*ΔNDA, and the lag dependant variable. Summing over the lags of ΔNDA and ΔNFA gives an offset coefficient of about 0.5 for the period prior to March 1995. Interestingly, for the period after the adoption of the crawling peg, the offset coefficient is fairly close to 1. However, the standard error of the coefficient is fairly large so the possibility that the offset coefficient is in absolute terms somewhat smaller, cannot be rejected. Moreover, equation (i) may overestimate the absolute value of the offset coefficient. For example, suppose there were other exogenous variables which led to attraction of capital from abroad, and the central bank attempted to sterilize the inflow by lowering its NDA, then in equation (i) the coefficient on ΔNDA would be overestimated. In fact, when the change in nominal output is introduced in equation (ii), the offset coefficient for the period prior to March 1995 drops to -0.38. However, again when summing over the lags, the long run offset coefficient for the period after March 1995 is close to -1. The final equation includes the current and first lag of foreign interest rates and nominal output. The interest rate variables are not significant; however, the order of the magnitude of the offset coefficients is similar to the other two equations.
With the above caveats in mind and not reading too much into the point estimates of the above equations, the results indicate that the effectiveness of sterilized intervention by the Central Bank has been substantially reduced since the preannounced exchange peg regime was adopted. The further liberalization of capital account controls since July 1996 will limit the scope for sterilized intervention even more.
References
Argy, V., and Kouri, P. (1974), “Sterilization policies and the volatility in international reserves”, In Aliber, R. Z. (ed.) National Monetary Policies and the International Financial System. University of Chicago Press, Chicago.
Darvas, V. (1996), “Exchange Rate Premia and the Credibility of the Crawling Peg Target Zone in Hungary”, CEPR Working Paper # 1307.
Kouri, P., and Porter, M. (1974), “International Capital Flows and Portfolio Equilibrium”, Journal of Political Economy. vol. 82, # 3, pp. 443-68.
Obstfeld, M. (1982), “Can we sterilize? Theory and Evidence”, American Economic Review, vol. 72, # 2, pp. 45-50.
Shadier, S, Carkovic, M, Bennett, A, and Kahn, R. (1993), “Recent Experiences with Surges in Capital Inflows” Occasional Paper no. 108, IMF.
Prepared by Juan Fernandez-Ansola and Reza Moghadam.
Darvas (1996) provides more detail on market segmentation.
This change in regulations made foreign currency deposits at the Central Bank less attractive for commercial banks. These deposits fell by over $1.5 billion in the period January-May 1996 (Table 1).
The monthly stocks in forints of each of these variables are converted into dollars at the end-of-period exchange rate, and dollar flows are calculated on this basis.
For Hungary, the nominal output series is constructed using the monthly industrial output series and the producer price index.
The interest differential appears in the structural model as a determinant of money supply. However, the domestic interest rate is substituted out in the reduced form. Here we use German interest rates as a proxy for foreign interest rates. In theory, i should be a combination of the foreign interest rate and the expected change in the exchange rate.
It proved difficult to find suitable “instruments” for ANDA in order to use the instrumental variables technique and avoid the problems of simultaneity.
Other specifications with variables such as the current account balance and the expected exchange rate were also estimated but these variables were either insignificant or had the wrong sign. However, in all cases ΔNDA was significant and had the correct sign.
