Conceptual Background and Methodology

Theory suggests that, in a small, open economy with a fixed exchange rate regime such as that of Oman, the scope of monetary policy independence is generally limited. Any attempt by the government to influence monetary aggregates, in a way that would result in a significant spread between the domestic interest rate and the interest rate on the peg currency, would lead to large foreign capital flows. These would neutralize the changes in the monetary aggregates and bring the domestic interest rate back to parity.¹ Sterilizing such flows can be feasible only in the short run, as it will eventually entice more capital flows unless the economic fundamentals change. In such an environment, the burden of stabilization falls, by and large, on fiscal policy, which, deprived of any active collaboration from monetary policy, has to be restrained in order to maintain price stability and to support the exchange rate.

In the short run, however, significant scope for monetary independence can exist, for various reasons. One is that although monetary policy cannot determine the size of the money supply in an open economy under a fixed exchange rate regime, it can still influence its composition; for example, it can determine the shares of domestic credit and of international reserves in the total. If international money is not a perfect substitute for national money in the short run, monetary policy will then have some independence (Poniachek, 1979). More important, various factors may reduce the efficiency of arbitrage in response to a divergence in interest rates. Such factors include delays in the response of capital flows to interest differentials because of lack of information, institutional factors, high transactions costs, or uncertainty regarding the feasibility of maintaining the peg in view of a deterioration in economic fundamentals. As a result, the government can exercise a reasonable degree of discretion over monetary policy in the short run.² This would, again in the short run, enable the government to coordinate fiscal and monetary policies more actively. For example, a larger fiscal deficit can be supported by higher interest rates, which will raise domestic financing and at the same time help control inflation. If these factors are at work in the economy, overwhelming inflows should not occur, and therefore there should be no need for large-scale sterilization, no loss of reserves, and no increase in foreign indebtedness.

The next subsection looks at recent trends in the interest rate on Omani treasury bills as compared with that on U.S. treasury bills. The existence of a significant spread between these rates is taken as an indicator of monetary independence. Obviously, such a spread may reflect not only Oman’s degree of monetary independence but also structural differences. For example, unlike the U.S. treasury hill rate, rates on Omani treasury bills are determined only in the primary market—that is, at their auction—since there is no secondary market for these bills. Also unlike U.S. treasury bills, Omani treasury bills are bought primarily by banks. Banks’ demand is not as price-elastic as that of private investors, since treasury bills are counted as reserve assets and thus help banks meet capital adequacy ratios. These considerations may qualify somewhat the results of this study. However, a large spread between the Omani and the U.S. treasury bill rate cannot be sustained for any length of time, given the openness of the Omani economy to international capital flows, unless the CBO has a certain degree of monetary policy independence.

The analysis also links developments in the domestic interest rate (and in the spread) to overall fiscal developments, in particular the domestic financing needs of the budget. Factors permitting monetary independence and restricting arbitrage in the past are also investigated.

An Empirical Investigation

Looking at the case of Oman during the period 1987–97, one can distinguish two main subperiods: the period before the financial liberalization of October 1993, and the period thereafter. We focus our analysis on the behavior of the spread between the interest rates on Omani and U.S. treasury bills and the factors behind that behavior. The two interest rates are plotted in the upper panel of Figure 6.1, and the spread between them is plotted in the lower panel. It appears that until late 1993 the government was able to exercise a considerable degree of monetary independence, as evidenced by the behavior of the interest rate on Omani treasury bills, which, for most of the period, did not track that on U.S. treasury bills. There was a significant spread between the two rates, and this spread changed from negative to positive during that period.

View Full Size

Omani and U.S. Interest Rates and Spreads

Sources: Data provided by the Omani authorities; IMF, International Financial Statistics, various issues.¹ Maturity of 90 days.

• Download Figure
• Download figure as PowerPoint slide

Two main factors acting in combination were responsible for this divergence in interest rates. First, interest rates in Oman were subject to various controls that hindered the free operation of the market. For example, interest rate ceilings were imposed on rial Omani deposits of the private sector as well as on commercial bank loans. These distortions affected the way the interest rate behaved so that it did not track the interest rate on the dollar. Second, there was apparently insufficient arbitrage to enforce interest rate parity, because of institutional factors, perceived exchange rate risk, and cost factors. For example, in March 1980 the CBO introduced a facility under which commercial banks could swap dollars at par for five days to three months.³ This facility was intended to improve banks’ ability to manage their own liquidity, but in effect it also encouraged arbitrage operations, as it led to a gain (loss) whenever U.S. interest rates were lower (higher) than that on the rial Omani. In June and July 1986, however, the facility was modified so that arbitrage operations would carry a charge, and this increased their cost. In addition, the modification made banks eligible for subsidies for losses arising from the interest rate differential; this also reduced the incentive for arbitrage.

The ability of resident banks to engage in arbitrage was also restricted by prudential regulations restricting lending and borrowing abroad. For example, effective March 1982, the net foreign exchange position of a resident bank could not exceed 40 percent of its capital and reserves. At the same time, since November 1981, commercial banks have been barred from both holding rial Omani deposits with, and advancing loans in rials to, nonresident banks and financial institutions. These restrictions, in effect, prevented arbitrage by nonresidents, as they were then not able to deal in Omani assets, especially because regulation and tax discrimination severely discouraged both direct and portfolio investment in Oman by foreigners.⁴ The scope for arbitrage was thus largely limited to arbitrage by resident individuals and to the maximum 40 percent of the relatively small capital and reserves of resident banks. The result of all this was a measure of independence of monetary policy in Oman—as evidenced from the data—during the period up to late 1993.

Looking at domestic interest rate behavior, one can relate these developments to the implicit short-term objective of monetary policy in Oman at that time, which was to raise sufficient domestic financing for the budget. For example, from late 1989 until late 1993, the spread between the interest rate on Omani treasury bills and that on U.S. treasury bills was largely determined by the amount of fiscal financing the government decided to raise, through treasury bills and the issuance of government development bonds at predetermined high interest rates. Indeed, during that period the interest rate spread moved in line with the volume of treasury bills outstanding. This period can be divided into two sub-periods: late 1989 to early 1991, and early 1991 to October 1993.

Developments from Late 1989 to Early 1991

During the first of these two subperiods, high oil prices had a positive impact both on the fiscal balance and on foreign exchange reserves. The government also took this opportunity to limit foreign financing of the budget in 1989 to only RO 46 million, down from about RO 200 million in 1988, and to partly replenish its assets abroad and reduce its foreign debt in 1990 (Table 6.1). So despite the improved overall financial situation, domestic financing increased because the government placed an emphasis on improving the net position of its foreign assets. As a result, domestic financing reached a historically high RO 244 million in 1989 before declining to about RO 190 million in 1990. This was also reflected in an increase in outstanding treasury bills from less than RO 5 million in mid-1989 to more than RO 50 million in early 1991.⁵

Table 6.1

Budget Financing Needs

(In millions of rials Omani except where noted otherwise)

Source: IMF, Government Finance Statistics, various issues.

^1/Negative numbers indicate deficits.

Table 6.1

Budget Financing Needs

(In millions of rials Omani except where noted otherwise)

Item		1985	1986	1987	1988	1989	1990	1991	1992	1993	1994	1995	1996
Overall fiscal balance 1		−364	−700	−146	−347	−290	−33	−284	−585	−511	−486	−468	−486
Financing		364	700	146	347	290	33	284	585	511	486	468	486
Financing abroad		374	595	92	197	46	−158	83	171	300	344	422	344
	of which:
	Changes in cash, deposits, etc.	300	379	144	125	1.1	−1.1	−2	67	317	216	423	216
	Net borrowing abroad	73	216	−52	72	35	−147	35	104	−17	123	−1	128
Domestic financing		10	−106	−54	−150	−244	−191	−202	−414	−211	−142	−46	−142
	of which:
	Monetary authorities	6	112	113	121	66	−32	102	105	151	−8	27	−8
	Domestic money banks	5	−20	−29	−24	85	12	−38	−6	8	−28	−44	−28
Memoranda:
Oil production (1990 = 100)		74	82	35	91	94	100	105	110	116	120	126	132
international oil price
	(dollars a barrel)	27	13	17	14	16	21	17	18	16	15	16	19
Foreign exchange reserves
	(millions of dollars)	1,043	915	1.347	1,004	1,748	2,228	2,607	2,394	1,873	1,715	1,899	2.031

Source: IMF, Government Finance Statistics, various issues.

^1/Negative numbers indicate deficits.

View Table

Table 6.1

Budget Financing Needs

(In millions of rials Omani except where noted otherwise)

Item		1985	1986	1987	1988	1989	1990	1991	1992	1993	1994	1995	1996
Overall fiscal balance 1		−364	−700	−146	−347	−290	−33	−284	−585	−511	−486	−468	−486
Financing		364	700	146	347	290	33	284	585	511	486	468	486
Financing abroad		374	595	92	197	46	−158	83	171	300	344	422	344
	of which:
	Changes in cash, deposits, etc.	300	379	144	125	1.1	−1.1	−2	67	317	216	423	216
	Net borrowing abroad	73	216	−52	72	35	−147	35	104	−17	123	−1	128
Domestic financing		10	−106	−54	−150	−244	−191	−202	−414	−211	−142	−46	−142
	of which:
	Monetary authorities	6	112	113	121	66	−32	102	105	151	−8	27	−8
	Domestic money banks	5	−20	−29	−24	85	12	−38	−6	8	−28	−44	−28
Memoranda:
Oil production (1990 = 100)		74	82	35	91	94	100	105	110	116	120	126	132
international oil price
	(dollars a barrel)	27	13	17	14	16	21	17	18	16	15	16	19
Foreign exchange reserves
	(millions of dollars)	1,043	915	1.347	1,004	1,748	2,228	2,607	2,394	1,873	1,715	1,899	2.031

Source: IMF, Government Finance Statistics, various issues.

^1/Negative numbers indicate deficits.

Despite the increase in outstanding treasury bills and other domestic financing, the government was able to maintain an interest rate that was lower than the interest rate on U.S. treasury bills, with a varying spread that reached 2.9 percentage points in August 1990. This was possible in part because, having increased from a low base, domestic financing needs remained manageable during that period.

Developments from Early 1991 to October 1993

Starting in 1991, however, the negative spread started to narrow; it turned positive by mid-1991. From then until October 1993 the interest rate on Omani treasury bills was higher than that on U.S. treasury bills by up to 2.6 percentage points (in October 1992). This was, again, a result of the government’s decision to restrain foreign financing and the depletion of foreign assets, which resulted in significantly increased domestic financing needs, especially as oil prices weakened and the fiscal deficit widened. During this period, domestic financing of the budget remained at the high level of about RO 200 million and reached a peak of RO 414 million in 1992, In addition to increased borrowing from the monetary authorities, outstanding treasury bills increased to RO 100 million by the end of 1991, RO 160 million by the end of 1992, and about RO 140 million in September 1993. (In October 1992 and July 1993 outstanding treasury bills were estimated at more than RO 200 million.) The government also launched development bonds in September 1991 at a high, administered interest rate to help raise needed budget financing. As a result, the net issuance of treasury bills and development bonds absorbed about 20 percent of national saving in 1991 and 1992. up from as low as 2 percent in 1990 (Table 6.2). This switch in policy, along with the small base of national saving (about 15 percent of GDP), created pressure in the domestic financial market, and this led to a rise in the interest rate. During 1993, however, the government reversed its policy and permitted more foreign financing. This alleviated some of the pressure in the domestic financial market, and as a result the spread started to narrow.

Table 6.2

Debt Instruments and National Saving

Sources: Data provided by the Omani authorities; IMF staff estimates.

Table 6.2

Debt Instruments and National Saving

Item	1987	1988	1989	1990	1991	1992	1993	1994	1995	1996	1997
		(in millions of rials Omani except where noted otherwise)
Outstanding treasury bills	36.0	12.0	10.0	33.0	104.0	162.0	111.0	45.0	20.0	66.7	56.1
Outstanding development bonds	—	—	—	—	40.2	113.9	151.7	221.2	227.6	229.2	236.3
Total	36.0	12.0	10.0	33.0	144.2	275.9	262.7	266.2	247.6	295.9	292.4
Net bill/bond financing	36.0	−24.0	−2.0	23.0	111.2	131.7	−13.2	3.5	−18.6	48.3	−3.5
In percent of national saving	4.3	−6.7	−0.3	2.2	18.7	21.5	−2.9	0.8	−4	0.8	—
					(In percent of GDP)
Gross domestic saving	33.8	23.8	28.9	32.0	23.3	25.5	22.5	24.0	23.5	27.5	27.3
Gross national saving	25.5	11.2	17.6	22.9	13.6	12.8	9.3	8.6	10.0	13.2	14.1

Sources: Data provided by the Omani authorities; IMF staff estimates.

View Table

Table 6.2

Debt Instruments and National Saving

Item	1987	1988	1989	1990	1991	1992	1993	1994	1995	1996	1997
		(in millions of rials Omani except where noted otherwise)
Outstanding treasury bills	36.0	12.0	10.0	33.0	104.0	162.0	111.0	45.0	20.0	66.7	56.1
Outstanding development bonds	—	—	—	—	40.2	113.9	151.7	221.2	227.6	229.2	236.3
Total	36.0	12.0	10.0	33.0	144.2	275.9	262.7	266.2	247.6	295.9	292.4
Net bill/bond financing	36.0	−24.0	−2.0	23.0	111.2	131.7	−13.2	3.5	−18.6	48.3	−3.5
In percent of national saving	4.3	−6.7	−0.3	2.2	18.7	21.5	−2.9	0.8	−4	0.8	—
					(In percent of GDP)
Gross domestic saving	33.8	23.8	28.9	32.0	23.3	25.5	22.5	24.0	23.5	27.5	27.3
Gross national saving	25.5	11.2	17.6	22.9	13.6	12.8	9.3	8.6	10.0	13.2	14.1

Sources: Data provided by the Omani authorities; IMF staff estimates.

Econometric Testing

To formally confirm the findings of the empirical analysis just presented, we tested econometrically for the existence of a significant cointegration between the interest rate on Omani treasury bills and that on U.S. treasury bills, using monthly data. We looked at the period 1987—96 and at three subperiods: 1987–89, 1989–93, and 1993–96. We started by establishing the order of integration of the variables.⁹ As expected, the stationarity tests confirmed that the variables are nonstationary. The first differences, however, turned out to be stationary, indicating that both variables are integrated of the first order. We then tested for cointegration during the whole period as well as during the three subperiods. As shown in Table 6.3, the results indicate that interest rates on Omani and U.S. treasury bills were not cointegrated during the period 1987–96. This confirms that the Omani interest rate did not closely follow the U.S. interest rate and implies a degree of monetary policy independence during the period as a whole. The results for the three subperiods, however, are somewhat different: Cointegration was found to exist during 1987–89 al a 95 percent confidence level, implying a rather limited degree of independence. No cointegration was found during 1989–93, implying a considerable degree of independence. Finally, during 1993–96 cointegration was observed at a degree of confidence of up to 99 percent, implying loss of monetary independence. These results are in line with the analytical and empirical discussions above and confirm their conclusions.

Table 6.3

Tests for Cointegration (Dickey-Fuller Statistic)

Source: IMF staff estimates.

^***/Indicates cointegration at 99 percent degree of confidence.

^**/Indicates cointegration at 95 percent degree of confidence.

^*/Indicates cointegration at 90 percent degree of confidence.

Table 6.3

Tests for Cointegration (Dickey-Fuller Statistic)

	Terms Included in the Residual Term Stationarity Test
	Neither trend nor constant term			Constant term only			Both trend and constant term
Period	0 lags	1 lag	2 lags	0 lags	1 lag	2 lags	0 lags	1 lag	2 lags
1987–96	−2.4821	−2.6020	−2.3089	−2.4821	−2,6020	−2.3089	−2.4822	−2.6023	−2.3149
1987–89	−3.6406**	−4.1831**	−3.1188	−3.6406**	−4.1832**	−3.1188	−3.7603	−4.2916**	−3.1964
1989–93	−1.9207	−1.9125	−1.6954	−1.9207	−1.9125	−1.6954	−2.7448	−2.6053	−2.4172
1993–96	−3.2860*	−3.7859**	−2.3875	−3.2860*	−3.7859**	−2.3875	−2.5657**	−6.0226***	−4.2107**

Source: IMF staff estimates.

^***/Indicates cointegration at 99 percent degree of confidence.

^**/Indicates cointegration at 95 percent degree of confidence.

^*/Indicates cointegration at 90 percent degree of confidence.

View Table

Table 6.3

Tests for Cointegration (Dickey-Fuller Statistic)

	Terms Included in the Residual Term Stationarity Test
	Neither trend nor constant term			Constant term only			Both trend and constant term
Period	0 lags	1 lag	2 lags	0 lags	1 lag	2 lags	0 lags	1 lag	2 lags
1987–96	−2.4821	−2.6020	−2.3089	−2.4821	−2,6020	−2.3089	−2.4822	−2.6023	−2.3149
1987–89	−3.6406**	−4.1831**	−3.1188	−3.6406**	−4.1832**	−3.1188	−3.7603	−4.2916**	−3.1964
1989–93	−1.9207	−1.9125	−1.6954	−1.9207	−1.9125	−1.6954	−2.7448	−2.6053	−2.4172
1993–96	−3.2860*	−3.7859**	−2.3875	−3.2860*	−3.7859**	−2.3875	−2.5657**	−6.0226***	−4.2107**

Source: IMF staff estimates.

^***/Indicates cointegration at 99 percent degree of confidence.

^**/Indicates cointegration at 95 percent degree of confidence.

^*/Indicates cointegration at 90 percent degree of confidence.

Conclusions

Past experience indicates that inefficiencies in the financial market, lack of sufficient financial instruments, and restricted foreign access to the domestic market contributed to a degree of monetary independence in Oman in the last decade. The domestic interest rate did not closely follow changes in the U.S. rate, and a spread was sustained whose magnitude corresponded to developments in the government’s domestic financing needs. The analysis presented in this section, however, indicates that, with the availability of more financial instruments, recent financial reform efforts, and increased globalization of financial markets, domestic financial intermediation has increased. This is reflected in the narrowing of the spread and the cointegration of the Omani interest rate with the U.S. rate in the 1993–96 subperiod. As a result, the scope for monetary policy independence in Oman is increasingly narrowing, and the role of monetary policy in the future will essentially be limited to liquidity management based on indirect monetary control. In this context it is fiscal policy that has to shoulder the central role in ensuring macroeconomic stability.