The Case for Adopting the Convertible Reserves System in Hong Kong

Pacific Economic Review, 3: 3 (1998) pp. 265–275

PERSPECTIVES ON PUBLIC POLICY

THE CASE FOR ADOPTING THE CONVERTIBLE RESERVESSYSTEM IN HONG KONG

TSANG SHU-KI* Hong Kong Baptist University, Hong Kong

Abstract. The East Asian currency crisis spread to Hong Kong in October 1997, touching offworldwide repercussions. Weaknesses were revealed in the operational mechanism of HongKong’s linked exchange rate system (the link). This paper casts the link as a peculiar currencyboard system in the proper historical and comparative contexts. After examining the conflictbetween the classical currency board (which relies on the two “automatic” stabilizers of specie-flow and cash arbitrage) and modern financial developments, it is pointed out that Hong Kongshould move forward to the convertible reserves system of Argentina, Estonia and Lithuania (theAEL model), and adopt its cashless/electronic arbitrage arrangement to defend the link.Moreover, it is necessary to balance the need for short-term stability and the consideration oflong-term flexibility. Any measure that increases the ultimate exit cost from the link must beassessed carefully.

© Blackwell Publishers Ltd 1998. 108 Cowley Road, Oxford OX4 1JF, UK and 350 Main Street,Malden, MA 02148, USA.

* Address for correspondence : Department of Economics, Hong Kong Baptist University,Kowloon Tong, Hong Kong.

1. THE LINK AND THE CLASSICAL CURRENCY BOARD

The Hong Kong dollar was pegged to the US dollar at the rate of 7.80 inOctober 1983, under the “linked exchange rate system,” or simply the “link.” Itwas supposed to be a currency board-type arrangement (Latter, 1993; NugÑee,1995; Sheng, 1995). The speculative attack on the Hong Kong dollar inOctober 1997, as a result of the “contagion effect” of the East Asian financialturmoil, and the subsequent continuing uncertainty threw open seriousquestions not only concerning the true nature of the link, but also about theviability of the currency board system as a genre in the modern era (Tsang,1997).

Currency boards originated in small British colonies in the 19th century.They reached their heyday in the 1940s, fell out of favor after World War II,and seem to have been enjoying a revival since the 1980s (Schwartz, 1993,Williamson, 1995).

In its classical version, a currency board issues currency notes and coins withfull foreign reserves backing at an officially sanctioned exchange rate.Theoretically, the exchange rate of deposit money is to be fixed at the officialrate because of two stabilizers. First is the specie-flow process. An outflow ofcapital, as a result of non-confidence in the system, would lead to a contractionof the money supply, push up the interest rates, and induce a counter-flow.Since the whole event would take place automatically and speedily, the

exchange rate can be fixed without government intervention (Hanke et al.,1993).

The problem with this is that higher interest rates would not necessarilyinduce a counter-flow, if the exchange rate is itself fluctuating and seen to beinsecure. Exchange rate risk would necessitate an interest rate premium, andhigher interest rates in turn might be regarded as a sign of weakness, evenleading to a vicious circle.

Hence the need for the second mechanism of the currency board: arbitragethat more immediately binds the exchange rate. Suppose the market exchangerate weakens from the official rate, people can first convert their bank depositsinto cash, go to the currency board to exchange the cash into foreign currencyat the stronger official rate, and then sell the foreign currency in the market,fetching an arbitrage profit. Since the board has foreign reserves that cover atleast 100 percent of cash in circulation, it can entertain all the requests ofexchanging cash into foreign currency. Similar to arbitrage under the old goldstandard where people shipped gold bullion across countries (Officer, 1989,1993), cash arbitrage appeals to the self-interest of market participants. Theselling pressure on the foreign currency will bring the market exchange rateback to the level of its official counterpart.

With these two automatic stabilizers that require no bureaucratic meddling, itis claimed that the currency board system has had a “perfect record” ofexchange rate stability. Steve Hanke states that “since the first currency boardwas installed in Mauritius in 1849, not one has suffered a successful specula-tive attack.”1

The attack in late October 1997 on the currency of Hong Kong, the onlyterritory in the East Asian region hosting a currency board system other thantiny Brunei which did not face any manifest troubles, threw some doubts onthe continuation of the “perfect record.” It appeared to many that Hong Konghad to struggle to protect the fixed exchange rate through pro-active actions bythe monetary authority, rather than purely relying on the two “self-adjusting”mechanisms. An application of Svensson’s (1991) “simplest test” on dailyexchange rate and interest rate data shows that the link was not “credible” inOctober 1997 and in January 1998 (when the Indonesian crisis touched offanother round of regional turbulence). Technical details are presented in theAppendix. Into 1998, Hong Kong had to cope with the unpalatable conse-quences of high interest rates and an unfolding recession. Some measuresshould in my view be implemented to shore up the robustness of the link and toalleviate the cost of defending it.

2. LESSONS FROM THE ATTACK ON THE LINK

What are the lessons to be learnt from the episode? The first is that the link wasa peculiar currency board system vulnerable to speculative attack. There was

© Blackwell Publishers Ltd 1998

266 TSANG SHU-KI

1 Steve H. Hanke, “The Solution: Autopilot for Hong Kong”, Asian Wall Street Journal, 30,October 1997.

actually no currency board, and notes were issued by three designated noteissuing banks (NIB),2 which alone could deal with the Exchange Fund at thefixed exchange rate. Notes-based arbitrage opportunities were therefore highlyrestricted, rendering one of the two stabilizers almost totally inoperative(Tsang, 1996a, b). Since the other stabilizer is in theory highly dependent onthe efficiency of arbitrage, it also did not function well. Hence the marketexchange rate persistently strayed from the official rate of 7.80, by an averageof slightly less than 1 percent since the inception of the link in October 1983.

In times of exogenous shocks or speculative attacks, Hong Kong’s monetaryauthority had to intervene in the foreign exchange market and manipulatebanking liquidity or interbank interest rates proactively.3 These were regardedby some as a departure from the classical currency board principles. However,there seemed to be few alternatives, unless the arbitrage mechanism could bestrengthened so that the market forces might be directly harnessed.

A controversy arose concerning whether and to what extent the Hong KongMonetary Authority (HKMA), the territory’s central bank, was responsible forthe unprecedented high interest rates during the speculative attack on the link inlate October 1997. Overnight interbank rates were up to 280 percent briefly onOctober 23. The HKMA argued that the system was on “auto-pilot” (Yam,1998), and that the Authority was just “sitting there passively” (Yam, 1998,paragraphs 21–2). Some critics did not accept that interpretation, as theHKMA openly warned banks early on October 23, that those which repeatedlyused the Liquidity Adjustment Facility (LAF) – the discount window – toborrow HK dollar funds would be penalized. This touched off a strong“announcement effect” that made banks scramble for funds. There were alsoreports of the HKMA intervening in the foreign exchange market, selling USdollars and buying Hong Kong dollars.

In a way, the arguing was immaterial to the real issue, as the deviations ofthe market rate from the official rate could have been the source of theproblem: in the heat of the East Asian crisis, speculators regarded the “non-fixity” as a sign of “insecurity” of the link, and decided to have a go at it. Ifthere had been no such deviations, because of, say, a more robust arbitragemechanism to be discussed below, speculators might not have come, or mightnot have been so aggressive. Then the question whether the phenomenalinterest rates were the result of “natural” specie-flow or the HKMA’sintervention was only academic.

A second lesson is that cash arbitrage is hopeless in defending a currencyboard regime in the modern context. Greenwood and Gressel (1988) detectedthe problem some time ago, and Tsang (1996a, b) tried to tackle it. In abanking system with fractional cash reserves, allowing depositors to convert


CONVERTIBLE RESERVES FOR HONG KONG 267

2 Coins are, however, issued by the Hong Kong government.3 For more details, see NugÑee (1995). NugÑee was with the Reserves Management Department

of the HKMA. Deputy Chief Executive Andrew Sheng also said, “We (the HKMA) use a wholerange of instruments in influencing the level of interbank liquidity to manage interbank interestrates, and consequently, maintain exchange rate stability” (Sheng, 1995, p. 61).

deposits into cash for the sake of doing arbitrage is a dangerous business, astable 1 on the Hong Kong dollar ratios can testify.

The relative size of the cash base has been contracting in Hong Kong. By theend of 1997, the Hong Kong dollar cash–deposit ratio in all “authorizedinstitutions” (including licensed banks, restricted-license banks, and deposit-taking companies) was less than 1 percent, while the territory-wide ratio ofnotes to deposits was only about 5.6 percent.4 It is difficult to imagine undersuch a situation that banks would want to facilitate arbitrage activities thatrequire the conversion of deposits into currency notes. Large conversion isequivalent to a bank run! Shipping gold bullion around has no similarhazardous consequence. This is a definite minus for the currency board systemvis-Ða-vis the gold standard. Actually, bank notes arbitrage was rarely observedin Hong Kong, and certainly not during the attack in October 1997.

Worse still, modern financial developments point irreversibly to a furtherdiminishing cash base, and the cashless society is widely predicted to arrivesome time in the 21st century. How then would a currency board be able to fixits exchange rate through cash arbitrage?

3. FROM CASH ARBITRAGE TO CONVERTIBLE RESERVES

In practice, Hanke et al. (1993, p. 5) observe that “in some cases” a currencyboard “issues deposits” fully backed by foreign reserves, and in theirrecommended “model constitution” for a currency board in Russia (Appendix


268 TSANG SHU-KI

4 This ratio, small enough, is still an exaggeration, because a significant amount, popularlyestimated at 20–30 percent, of Hong Kong dollar notes have been circulating in the southern partof China.

Table 1. Year-end HK$ cash–deposit ratios in Hong Kong

Ratio of notes and coinsNotes in circulation (i.e. cash) held by all

over total HK$ deposits “authorized institutions”Year for whole economy (%) over total HK$ deposits (%)

1985 7.60 1.101986 7.55 1.141987 8.14 1.211988 8.15 1.081989 8.27 1.221990 7.86 1.071991 7.70 1.091992 8.51 1.311993 8.03 0.991994 7.32 1.021995 6.63 0.881996 5.90 0.761997 5.59 0.84

Sources : Census and Statistics Department, Annual Digest of Statistics; HongKong Monetary Authority, Monthly Statistical Bulletin.

A), they also allow for such a possibility. Since they strongly argue against“central banking” and regard the currency board system as a superior, self-adjusting alternative that does away with “discretionary power” (see alsoHanke and Schuler, 1994), it is not clear why a currency board should acceptany deposits from commercial banks and then cover them with foreigncurrency. In any case, in their treatise, they have not spelt out the implicationsfor arbitrage if banks do place deposits with the currency board.

In an IMF survey, BaliÓno and Enoch (1997) look at modern “currency boardarrangements” (CBAs) in general. In quite a few of the CBAs in existence(which only total about a dozen), there is ironically a central bank. They notethat in several cases (including Argentina, Estonia and Lithuania), the centralbank takes deposits from commercial banks and its foreign exchange reservesextend beyond the amount of notes and coins in circulation to cover thosedeposit liabilities. Unfortunately, BaliÓno and Enoch have not highlighted theoperational differences between a CBA that guarantees the convertibility ofonly notes and coins and one that guarantees the convertibility of the total ofbanking reserves (or the monetary base in the modern context) with it. In fact,there are very significant implications, particularly with regard to arbitrageefficiency and the stability of the exchange rate.

Tsang (1996a, b) singles out Argentina, Estonia and Lithuania for detailedinvestigation, and finds that their model (the “AEL model”) overcomes theproblems of cash-based arbitrage. These three countries began a currency-board type system in 1991, 1992 and 1994, respectively (BaliÓno and Enoch,1997; Bennett, 1993, 1994). Though latecomers compared with Hong Kong,they use an improved arrangement which shows a much higher degree ofarbitrage efficiency and exchange rate stability. The spot rate in the foreignexchange market has been invariably quoted around the official rate, despiterecent political and economic turbulence.

In these three countries, banks have an account with the central bank, inwhich deposit reserves as well as other balances are kept. The central bankguarantees the full convertibility of these bank balances, at the fixed exchangerate. This setup bypasses the problem of moving cash around for arbitrage,which is a core defense mechanism in the classical currency board.

Let us look at a hypothetical example in a country where the domesticcurrency, called peso, is pegged to the US dollar at parity under the convertiblereserves scheme similar to the AEL model. Every bank is then obliged to quotethe official rate of 1 peso per US dollar. Suppose there is a deviant, Bank A,which quotes an exchange rate of 1.1 peso. Bank B can sell US$1 million toBank A for 1.1 million pesos, asking A to transfer the pesos to B’s account atthe central bank. (B would of course transfer US$1 million to A’s accountthere.) On demand, the central bank would convert the pesos into US$1.1million for Bank B, which then fetches an arbitrage profit of US$100 000. A,on the other hand, suffers a loss of 100 000 pesos as its US$1 million at thecentral bank can only be turned into 1 million pesos. If Bank A remainsunrepentant, every other bank would be jumping on it.

No cash movements are involved, as the central bank plays the role of



clearing the arbitrage transactions between the two banks (Tsang, 1997). Notethat after settlement, the central bank’s foreign reserves will be reduced byUS$100 000. In other words, the central bank risks losing reserves if acommercial bank like A rebels against the peg. However, that loss is matchedby a correspondent shrinkage (100 000 pesos) in Bank A’s balance sheet. Sincethe deal is settled by accounting transfers through telephone calls and electronicmeans, the transaction cost is reduced to an absolute minimum. This arrange-ment for “cashless” or “electronic” arbitrage is obviously superior to the cash-based arbitrage mechanism under the classical currency board. Its efficiency isalso higher than arbitrage under the old gold standard (Officer, 1989).

In reality, under the convertible reserves system, no banks would dare todeviate in quoting exchange rates. All commercial banks are bound by the rule ofthe game to quote the official exchange rate, within a very narrow buying andselling spread that truly reflects petty transaction cost, lest they will be hit by theirmarket rivals. Therefore, no actual arbitrage needs to take place, and the centralbank is in no fear of losing foreign reserves.5 With this improved form of CBA,Argentina, Estonia and Lithuania have been able to literally fix their spotexchange rates despite serious economic or political instability (Tsang, 1997).

4. CONVERTIBLE RESERVES SYSTEM FOR HONG KONG

Hong Kong can adopt the convertible reserves system to rein in the marketexchange rate and as a much firmer defense against currency attacks. A centralbank, the HKMA, came into existence on April, 1 1993. In late 1996, a realtime gross settlement (RTGS) system was instituted, replacing the traditionalarrangement where a commercial bank, the Hongkong Bank, served as theclearing bank for the financial sector. The RTGS is an accounting mechanismbetween the HKMA and all commercial banks, which allows the former totransact directly with the latter.

On the basis of the RTGS infrastructure, it is a modest step to institute asystem under which every bank has a direct reserve account with the HKMA.One top of the normal clearing balances, the HKMA could ask each bank tosubmit an equivalent amount of US dollar to it for obtaining notes from theNIBs.6 Concurrently or alternatively, the HKMA might impose a depositreserve requirement on the banks. To overcome resistance from the bankingsector, the ratio, which could be interpreted as a “financial tax,” should be assmall as possible. Near-market interest rates could also be paid on those reservedeposits. The idea is not to tax the banks, but to ensure that there is suitableliquidity in the reserve account to minimize any possible impact on the interestrate. As explained above, if the system works, no actual arbitrage needs to take


270 TSANG SHU-KI

5 This game-theoretic set-up must be the basis of the gold standard and the currency boardsystem, which rely on the self-interested activities of market participants to cancel each other out,thereby holding the exchange rate. See Tsang (1984).

6 This arrangement would be similar to the pre-1994 situation, although commercial banks thenpassed the US$ reserves to the NIBs rather than to any central monetary authority individually.See Tsang (1996a).

place and the spot exchange rate will be fixed around the official rate. In thatsituation of benign equilibrium, the deposit reserve ratio could even approachzero.

At US$92.8 billion, Hong Kong’s foreign exchange reserves at the end of1997 were the third largest in the world (after Japan and China). Theyrepresented more than seven times of the amount of HK$ notes and coins incirculation and over 40 percent of the HK$ M3, and were much larger, inabsolute and relative sizes, than those of the AEL countries. The HKMAshould have little problem in hosting the arbitrage game among financialinstitutions. Adopting the AEL model would indeed require a lower degree ofactivism on the part of the HKMA. The Hong Kong economy should face lessunpalatable results in the case of a return of the speculators, as the latter cansee that all banks are bound to quote around the rate of 7.80. In other words,they would be fighting the whole banking system, rather than the HKMA, intheir bid to unsettle the link.

5. EFFICIENCY RISK, SYSTEMIC RISK AND EXIT COST

Of course, although the spot exchange rate is “fixed,” the forward rate is not.Local interest rates could still be higher than those of the foreign counterpart,along with weak forward exchange rates. The reason might be that marketparticipants are not sure whether the convertible reserves system could reallyfix the spot rate. In other words, there is an “efficiency risk” and they demand arisk premium. However, over time the fixation of the spot rate should lead tothe return of confidence, and convergence in interest rates and exchange rateswould take place as people engage in interest arbitrage. That has beenoccurring in Argentina, Estonia and Lithuania, as analysed by BaliÓno andEnoch (1997, Appendix I).

That convergence has, however, not been perfect in the three countries: localinterest rates have still been higher than those of the US dollar (to which theArgentine peso and the Lithuanian litas are pegged) and the German Mark (towhich the Estonian kroon is linked). This is due to the existence of “systemicrisk.” In other words, although market participants observe the fixity of thespot rate, they are not sure that the “perfect” system that is working so wellwill not be abandoned in the future, perhaps not because it is defective, butbecause of other political and economic factors. Commentators familiar withthe situations in these three countries understand why some people might benervous, justifiably or otherwise, about the possibility of coup d’Ñetat orexternal invasion. Moreover, no matter how good it is in anchoring theexchange rate, whether a fixed rate regime is optimal for the economy is also acontroversial issue.

Governments in the three countries have tried to contain the marketperception of systemic risk by legal means (BaliÓno and Enoch, 1997; Tsang,1996a, b). In Argentina and Estonia, it was enshrined in an act of the Congressand the Parliament respectively that the central bank can only revalue but notdevalue the exchange rate. In Lithuania, according to the Law on Litas



Credibility, the exchange rate could only be changed by the Bank of Lithuania,in consultation with the government, under extraordinary circumstances.

In the case that Hong Kong adopts the convertible reserves system, theefficiency risk should be eliminated rather quickly. Interest rate convergencecould unfold at a faster pace than that observed in those three countries. As tothe systemic risk, Hong Kong’s political and economic situations are far morestable than theirs: no one here would seriously fear the possibility of coupd’Ñetat or invasion, although re-pegging or re-floating out of optimalityconsiderations can never be entirely ruled out.

Enshrining the link in law may further reduce systemic risk and lead to aneven higher degree of interest rate and spot-forward exchange rate conver-gence. In any case, the convergence will never be perfect as the law itself isstill open to some residual doubts. Moreover, it will actually increase the “exitcost,” in case it is deemed optimal to abolish the fixed rate regime and re-floatthe Hong Kong dollar in the future. Alternatively, when all conditions maturesome time in the 21st century, it may become advisable to peg the localcurrency to the then freely convertible Chinese currency, the Renminbi, insteadof sticking with the US dollar.7

The same problem of exit cost also applies to any scheme under which theHKMA issues insurance instruments (e.g. put options) to market participants tofoster confidence that the link rate will not be changed, as well as the proposalof “dollarization,” i.e. replacing the Hong Kong dollar with the US dollar.Compensations might be huge for a permanent insurance scheme, and the re-introduction of a Hong Kong currency could be a troublesome affair after“dollarization.” The Hong Kong government needs to strike a balance betweenthe conflicting considerations of “eliminating” the systemic risk for the linkand reducing the cost of exiting from it.

APPENDIX: CREDIBILITY TESTS ON THE HONG KONG LINK

Svensson’s (1991) “simplest test” of target zone credibility is performed on thelink for the period around the speculative attack in late October 1997. Theassumption is that, because of arbitrage imperfection, the market rate woulddeviate to a certain extent from the official rate of 7.80. Nevertheless, thesystem still holds as few would doubt its continuation. Svensson’s test firstcomputes the rate of return of a foreign currency investment for τ months, Rτ

t,given a band within which the central bank defends the exchange rate. Hencethere are an upper bound (KRτ

t) and a lower bound (Rτt) of the rate of return. The

upper bound is calculated as:

KRτt = (1 + it*

τ)(MS/St)12/τ − 1


272 TSANG SHU-KI

7 It is interesting to note that both Estonia and Lithuania have started the process of applying tojoin the European Economic and Monetary Union (EMU). A pre-requisite for eventualmembership means that they have to give up their CBAs, no matter how perfect they are, becausethe EMU hosts a central banking system. Lithuania has actually announced that they are exitingfrom their CBA in stages during 1997–99. See Bank of Lithuania, The Monetary PolicyProgramme for 1997–99, January 16, 1997. This shows the importance of considering the issueof exit cost for CBAs.

where it*τ represents the foreign interest rate in time t for a τ-month loan or

investment, St the spot exchange rate (expressed as the ratio of the domesticcurrency per unit of foreign currency), and MS the upper bound of the exchangerate (i.e. the limit of depreciation allowed). Likewise, the lower bound of therate of return is given by

Rτt = (1 + it*

τ)(S/St)12/τ − 1.

Under the assumption of there being no arbitrage, a completely credibleexchange rate implies that the domestic interest rate it must lie within the targetzone of Rτ

t, i.e.

Rτt ≤ it ≤ KRτ

t

If the domestic interest rate moves above the upper bound, the no-arbitrageassumption implies that the exchange rate regime cannot be completelycredible as market participants collectively perceive a risk of devaluation. Asymmetrical 1 percent limit on either side of 7.80 is used as the “target zone”for the link, on the basis that the government did seem alerted when thedeviation from the official rate approached 1 percent. Hence MS is 7.878 and S is7.722 for the zone.

Another variant of Svensson’s (1991) test is based on the assumption ofuncovered interest parity. Let us look at the following equation:

EtSt + τ = St[(1 + itτ)/ (1 + it*

τ) ]τ/12

where EtSt + τ is the expected value in time t of the ruling exchange rate inmonth (t + τ). The right-hand side of the equation is the annualized interestdifferential between the domestic and foreign interest rates, adjusted for thematurity period of τ months. One can check whether the expected exchange



Figure 1. Svensson’s “simplest test” on the link using daily data (no arbitrage)

rate (determined by the interest differential) ever moved outside the “targetzone,” i.e. MS and S.

I have implemented the two variants of Svensson’s test using daily closingdata running from January 1997 to February 1998. Figures 1 and 2 show theresults of the two tests. The three-month interbank offer rate in Hong Kong(HIBOR) is used as it

3 and the three-month Eurodollar interest rate is employedas a proxy for it*

τ. The results indicate that the link was not “credible” in faceof the currency attack, given a 1 percent target zone. Credibility was severelyshaken in October 1997 and less so in January 1998. It appears that marketparticipants suspected that the Hong Kong dollar would devalue. Of course, acaveat is that a 1 percent “zone” was not official. Moreover, target zoning is intheory not the same as currency board economics.

REFERENCES

BaliÓno, T. and C. Enoch (1997) “Currency Board Arrangements: Issues and Experiences,” IMFOccasional Paper, No. 151, August.

Bennett, A. G. G. (1993) “The Operation of the Estonian Currency Board,” IMF Staff Papers40(2), June, 451–70.

—— (1994) “Currency Boards: Issues and Experiences,” IMF Papers on Policy Analysis andAssessment, PPAA/94/18.

Greenwood, J. and D. Gressel (1988) “How to Tighten the Linked Rate Mechanism,” AsianMonetary Monitor, January–February, 2–13.

Hanke, S. H. and K. Schuler (1994) Currency Boards for Developing Countries, San Francisco:International Center for Economic Research.

Hanke, S. H., L. Jonung and K. Schuler (1993) Russian Currency and Finance, London:Routledge.

Hong Kong Monetary Authority (1994) The Practice of Central Banking in Hong Kong.Latter, A. (1993) “The Currency Board Approach to Monetary Policy – from Africa to Argentina


274 TSANG SHU-KI

Figure 2. Svensson’s “simplest test” on the link using daily data (uncoveredinterest parity)

and Estonia, via Hong Kong,” in Hong Kong Monetary Authority, Monetary Management inHong Kong, proceedings of the Seminar on Monetary Management, 26–43.

NugÑee, J. (1995) “A Brief History of the Exchange Fund,” Quarterly Bulletin, Hong KongMonetary Authority, May, 1–17.

Officer, L. H. (1989) “The Remarkable Efficiency of the Dollar–Sterling Gold Standard,1890–1906,” Journal of Economic History 49(1), 1–41.

—— (1993) “Gold-point Arbitrage and Uncovered Interest Arbitrage under the 1925–1931Dollar–Sterling Gold Standard,” Explorations in Economic History 30(1), 98–127.

Schwartz, A. J. (1993) “Currency Boards: Their Past, Present and Possible Future Role,”Carnegie–Rochester Conference on Public Policy 39, 147–93.

Sheng, A. (1995) “The Linked Exchange Rate System: Review and Prospects,” QuarterlyBulletin, Hong Kong Monetary Authority, May, 54–61.

Svensson, L. E. O. (1991) “The Simplest Test of Target Zone Credibility,” IMF Staff Papers38(3), 655–665.

Tsang, S. K. (1984) “On the Cash-based Fixed Exchange Rate System,” in The Pearl in theMouth of the Dragon: Collected Essays (in Chinese), Hong Kong: Wide Angle Press,179–201.

—— (1996a) “The Linked Rate System: through 1997 and into the 21st Century,” in NgawMeekau and Li Si-ming (eds), The Other Hong Kong Report 1996, Hong Kong: TheChinese University Press, chapter 11.

—— (1996b) A Study of the Linked Exchange Rate System and Policy Options for Hong Kong, areport commissioned by the Hong Kong Policy Research Institute, October.

—— (1997) “Currency Board the Answer to Rate Stability,” Hong Kong Standard, 31 October.Williamson, J. (1995) What Role for Currency Boards? Washington, DC: Institute for

International Economics.Yam, J. (1998) “Hong Kong: Financing Asia’s Development,” Keynote Address, Hong Kong

Development Council Financial Roadshow in Tokyo, March 3.



Documents

The Case for Adopting the Convertible Reserves System in Hong Kong