Alternative monetary policy instruments
Since widespread direct controls on financial intermediation and foreign exchange transactions were removed in 1984/85, monetary policy has worked through the ability of the Reserve Bank to influence interest rates. In the absence of direct administrative controls affecting the supply of, inter alia, credit and foreign exchange, modern monetary policy here and abroad works indirectly, relying on the ability of interest rate changes to affect economy-wide demand.
Relying on interest rates to influence demand, spending, and inflation has sometimes been distinctly uncomfortable. A concrete example was the marked, and sustained, rise in the exchange rate during the mid-1990s, at a time when the strongest inflationary pressures were clearly concentrated in the domestic non-tradables sectors, rather than in the tradables sector. We looked at whether there were any supplementary measures that, if adopted, might ease the apparent imbalances or smooth the adjustment. We concluded that none of the options would have materially eased the cyclical imbalances, and in many cases any such gains would have come at the cost of significant ongoing costs to the efficient operation of New Zealand's economy and financial markets.
A brief review of New Zealand's experience
1. For most of New Zealand's post-war history, monetary policy operated primarily through administrative instruments. The exchange rate was largely fixed, and for most of the period there was relatively limited scope for most interest rates to move. Tight controls largely eliminated short-term capital flows, while exchange controls and import licences were used to affect spending on imports. Closer to the heart of monetary policy, reserve ratios were imposed on the increasingly wide range of financial institutions. Raising a ratio was intended to limit the funds banks had available to lend. As they were typically unable to generate additional funds by competing for deposits through higher interest rates, banks were effectively supply-constrained. Rather than incur the penalties involved in failing to meet the ratio requirements, banks were forced to ration directly the available credit, in turn influencing domestic spending and economic activity.1 To a very significant extent, monetary policy operated by shortening or lengthening the queues for credit, with the tap turned off when inflationary pressures appeared to be particularly threatening.
2. In its earlier years, in particular, this system provided effective monetary control, without big fluctuations in either interest or exchange rates (for example, the nominal exchange rate was not adjusted between 1949 and 1967). A substantial administrative apparatus combined with relatively undeveloped financial markets helped make that possible. And yet, the controls themselves created the incentives for the rapid development of an increasingly wide range of new financial institutions and markets.
3. Turning off the bank credit tap by administrative fiat was a fast and technically efficient tool, but those who missed out on bank credit were often willing to pay well above the prevailing bank interest rates if they could find a lender operating beyond the boundaries of the control regime. Interest rates themselves were gradually liberalised over the years, though not without temporary reversals. But whenever administrative restrictions such as ratios weighed heavily on banks, other institutions would pick up the slack to some extent, diminishing the importance of banks in the overall process of financial intermediation.
4. As financial markets developed, it became progressively more difficult and costly to maintain a regime of direct controls. Increasingly, the direct controls themselves were working mostly through the effect they had on interest rates. Sometimes that meant interest rates across the economy, but in more intensively regulated periods might mean primarily the interest rates in the relatively less regulated markets (such as those for commercial bills, or solicitors' mortgages), to which the marginal borrowers turned for credit. The system made the allocation of credit a complex and inefficient process, biasing the flow of credit towards longstanding and existing borrowers, and towards financing investment in bricks and mortar, and away from newer or more adventurous propositions. That, in a sense, entrenched a slow pace of adjustment to changing economic circumstances. Financial institutions were able to earn a very comfortable return on their lending business, but at the cost of stifling innovation in the core financial sector. In addition to the increasingly obvious efficiency costs, equity concerns also suggested a case for change: the control-oriented system typically impinged most heavily on smaller savers and borrowers.
The nature of the concern
5. In deregulated financial markets, the notion of queuing for credit is largely alien. Whether in booms or slumps, banks compete to encourage us to borrow. Not all borrowers or projects will be regarded as creditworthy, but loans are not denied simply because banks do not have the funds to lend. Similarly unfamiliar to participants in a liberalised financial system is the notion of limited access to foreign exchange, whether for trade or investment purposes, or of foreigners' having only limited access to New Zealand dollars. Monetary policy operates instead primarily by influencing firms', households', and investors' demand for New Zealand dollars.
6. For the most part, that influence doesn't work overnight, and is often quite frustratingly slow. That is partly because interest rates tend to be rising at just the time that sentiment about future prospects is becoming increasingly optimistic. Against that sort of backdrop, most potential borrowers will regard an additional 25 basis points as a mere trifle, by contrast with the expected gains to be had from going ahead with the borrowing and spending they had planned.
7. Meanwhile, sustained higher interest rates will often attract foreign buyers of New Zealand dollars, pushing up the exchange rate, and shifting some - sometimes much - of the adjustment burden onto the export sector, regardless of whether any inflation pressures were apparent in that sector. The higher exchange rate that results is likely to mean that interest rates rise less than would otherwise have been the case. The effect can, of course, work in reverse: a buoyant export sector is likely to result in a higher exchange rate and higher interest rates, shifting some of the burden of the adjustment onto sectors that perhaps had benefited scarcely at all from the export boom.
8. With only one instrument to adjust - interest rates - these imbalances are frustratingly inevitable: a single interest rate means one interest rate for all regions and one interest rate for all sectors. But are there any alternatives that might supplement, or speed up the effectiveness of, the interest rate instrument? Or which might limit the spillover from changes in domestic interest rates into the exchange rate? Or simply limit the extent of the exchange rate cycles that seem to have become such an established feature of a market-oriented financial system? Is there anything that should be put back in the tool kit, to enhance our - very imperfect - ability to influence credit creation and manage the amplitude of cycles in monetary conditions?
Is there a role for quantitative tools?
9. A superficially obvious option is the possibility that reserve ratios (or similar direct quantitative controls) might still have a role to play. In 1985, New Zealand was the first country to abandon reserve ratios completely, and we are still among the minority in not having a ratio system at all. Some have suggested the use of ratios might influence the supply of money and credit more directly, in turn reducing the need to rely on fluctuations in interest rates to manage the demand for credit. Advocates of such measures have highlighted the very dramatic growth in credit during the 1990s. If ratios added something, they might help dampen unwelcome fluctuations in monetary conditions.
10. In some circumstances, reserve ratios (or other quantitative tools) can have a material impact independent of an associated change in interest rates: as noted above, they did so here in previous decades. The most obvious circumstance is when interest rates themselves are regulated. When interest rates are regulated, discretionary credit rationing is the only mechanism through which bank balance sheets can adjust. If wholesale financial markets were either absent or ill-developed - so that the supply of deposits was largely inelastic in the short-term - a quantitative tool like a reserve ratio might help limit overall credit - and aggregate demand - growth with a more-muted interest rate effect than otherwise.
11. Neither of those circumstances prevail in New Zealand today. An active wholesale market integrated with world financial markets means that banks can fund almost any lending demand, and do not need to engage in rationing (any more than, say, a supermarket does in normal circumstances). So, in New Zealand - and other developed countries - an increase in a reserve ratio would work as follows: each bank individually would need to acquire more cash, and the expression of that demand in the market would bid up the price of cash (the interest rate). Higher rates would attract additional deposits and - as those higher costs were passed on through a higher price of credit - would discourage some borrowers. But there is no mechanism through which banks would be forced (or even encouraged) to make administrative credit rationing decisions (based on the availability of funds to lend) over and above reliance on the price mechanism alone.2 A "very large" change in a ratio might prompt some rationing, but it would also prompt a very dramatic interest rate response.
12. A well-designed ratio system is likely to do little harm (although there are always costs in managing such an administrative system). However, once interest rates have been liberalised and a wholesale deposit market has developed, the rationale for a ratio-based system disappears: a reserve ratio would be redundant elaboration of the control system. Markets more generally do not usually voluntarily use administrative rationing to supplement the price mechanism. There are exceptions, but there is no reason to expect that imposing a reserve ratio system would make financial markets one of those exceptions.3
13. For these sorts of reasons, the rise of liberalised financial markets has led to quantitative mechanisms fading from prominence in policymakers' tool kits throughout the world. No developed country now uses ratios actively as part of monetary policy, or relies materially on other quantitative mechanisms to manage credit creation. The new European Central Bank, for example, after much debate, adopted a minimum reserves ratio, but there is no sign that this ratio will be used for monetary policy purposes. Similarly, the United States has a reserve ratio system in place but it has not played an important role in monetary policy implementation for many years.
14. In liberalised markets, in which banks can readily raise additional funds in wholesale markets, there is also little likelihood that open market operations or other aspects of public debt policy might help directly influence the availability of credit, limiting the need to use the interest rate tool. Transactions such as these would have an effect, nowadays, only if they could alter interest rates. At the margin, changing the structure of the government's debt sales programme may make it possible to slightly influence the slope of the yield curve, but such adjustments would be unlikely to affect the average level of interest rates facing investors or borrowers.
15. Developments in world financial markets over the late 1990s have renewed a debate as to whether capital controls constitute a sensible response to surges in capital inflows, "throwing sand in the wheels" of international markets to limit, to some extent, short-term cyclical capital inflows.
16. The strong capital inflows into New Zealand in the mid 1990s reduced the extent of the rise in interest rates. Without the possibility of capital flows, interest rates would have been much higher, leaning directly against the domestic inflationary pressures. Could "sand in the wheels" measures have reduced the volume of capital inflows, allowing interest rates to be higher? Linked to this, might controls have helped dampen the cyclical fluctuations in the real exchange rate, themselves a symptom of strong capital flows?
17. At the beginning of the 1990s, Chile faced a surge in short-term capital inflows. In response, for most of the 1990s, Chile ran a set of capital controls designed to influence the size and composition of such inflows. More recently, at the height of the Asian crisis, Malaysia also moved to clamp down on capital flows. Chile's experience in trying to manage short-term capital inflows provides an important case study. The Reserve Bank has put some effort into understanding the impact of the Chilean control system
18. Aiming to limit an appreciation in the exchange rate, the Chilean authorities believed that the first-best solution was more fiscal tightening. However, the fiscal surplus was already high - in the order of 2 per cent of GDP - and politically impossible to raise further. The monetary authorities tried removing the existing restrictions on capital outflows, but this created the perverse reaction of even more inflows entering the country.
19. Hence, the goal of capital controls was to introduce a "wedge" between domestic and world interest rates. That is, the Chilean monetary authorities wanted to be able to raise domestic interest rates, if overall inflation pressures warranted that, without incurring the same rise in the volume of capital inflows that would have occurred in a totally liberalised system.
20. Two elements of the Chilean capital controls scheme in the 1990s are relevant to this discussion.
21. The first was an Unremunerated Reserve Requirement (URR) on foreign loans and fixed-income securities. The URR was a one-year compulsory non-interest earning deposit at the central bank of a share of inflows of short-term and debt-creating capital. The requirement to hold the deposit for one year meant that the financial burden diminished as the term of the investment lengthened. The second element of the capital controls was the imposition of a minimum holding period of one year for foreign direct investment.
22. Between 1991 and 1997, the rate of the URR was increased and its coverage extended. From an initial level of 20 per cent, the rate was increased in 1992 to 30 per cent and coverage was extended. Nonetheless, coverage of the URR was partial: some potentially volatile short-run flows, such as trade credits and foreign exchange transactions in the informal or grey market, were excluded. Due to a sharp fall in capital inflows following the Asian crisis, the URR was reduced to 10 per cent in July 1998, and then to 0 per cent in September 1998.
23. Both within Chile and abroad, there is wide disagreement about the impact of the controls. This is partly because the controls changed over time (particularly as the private sector engaged in activity to avoid them), and partly because there were other important policy and regulatory factors that are difficult to adequately control for in doing an assessment.
24. Unsurprisingly, Chile's reserve requirement policy created incentives to find loopholes or to simply evade the reserve requirement. This produced a cat-and-mouse game between market participants and the central bank, with the latter attempting - and often succeeding, at least temporarily - in closing loopholes, cracking down on evasion, and gradually broadening the reserve requirement. In assessing the Chilean experience, it appears the capital controls:
- were not initially evaded, but subsequently the effectiveness of the controls was gradually eroded;
- were able to create a positive differential between domestic and foreign interest rates, although the true effect was probably small and relatively transitory;
- probably did not reduce the overall amount of capital inflows but probably did reduce the share of short-term flows and raise long-term flows; and
- probably did not substantially mitigate a real exchange rate appreciation.
25. And, of course, the controls created a number of costs. First, there were resource and opportunity costs. Resources were diverted to administration and evasion, while the controls retarded the integration of the economy with international markets (in one sense, that was the point of the controls, but it also meant missing out on the benefits of integration). Second, capital controls of this sort tend to be inequitable. In Chile, for example, the reserve requirement tended to penalise small firms relative to large ones.
26. It is likely that introducing such controls here would have come at a high cost to New Zealand's standing in world financial markets. Particularly in the transition, that might well have meant considerably less upward pressure on the exchange rate than otherwise, and a change in the "mix" of monetary conditions in the direction we had often hankered after. Such apparent gains would have had to be weighed against the longer-term costs of making it harder, or more costly, for New Zealand borrowers to access world capital market: across all stages of the cycle, and not just when New Zealand was wanting to lean strongly against domestic inflation.
27. New Zealand's well-developed financial markets, especially for derivative products, would be likely to have led to successful attempts to circumvent the intended effects of the controls. As with all regulatory policies of this sort, the effectiveness of the measures would probably be related to the effort the authorities are prepared to devote to ensuring comprehensive coverage and effective enforcement. In our view, Chilean-style controls would have been very difficult to implement in New Zealand with any degree of enduring effectiveness, at least without incurring very substantial supervisory and compliance costs.
Other tax or regulatory interventions?
28. Throughout both the disinflationary period of the late 1980s, and when the pressures created by a tight monetary policy were "biting" in the mid-1990s, the Reserve Bank spent time considering whether there were other policy measures the government could adopt that might alleviate some of the pressures. Ideally, these would have been measures that made sense in their own right, contributing to maximising New Zealand's long-term economic growth potential. In the late 1980s, this list included substantially reducing the then-high fiscal deficits, and cuts to New Zealand's then high levels of trade protection. Our analysis, and experience abroad, suggested that measures in these areas would ease the degree of monetary policy pressure needed to keep inflation under control, and perhaps mitigate the extent of cyclical exchange rate peaks.
29. By the mid 1990s, the options were less obvious. To take fiscal policy as one example, New Zealand ran fiscal surpluses from 1993/94 onwards. Whatever the cyclical case might have been for a tighter fiscal policy, it was hard by then to argue that longer-term growth or sustainability considerations required higher surpluses than those projected at the time.
30. However, recognising that marked cyclical fluctuations in financial prices can themselves be costly, some effort was devoted to exploring whether there were other policy initiatives available that might ameliorate the cyclical problems while either delivering improved microeconomic efficiency, or at least without creating material longer-term costs. Some of these might have involved using our prudential supervisory powers, and are dealt with in another paper in this series. [See "Prudential and monetary policy".]
31. Other ideas involved the tax system. The Reserve Bank could not have imposed such measures, but they could have represented options if the government itself come to the conclusion that the cyclical fluctuations, and the imbalance in the mix of monetary conditions, had become too onerous.
32. Options included several possible "sand in the wheels" strategies, such as the reimposition of full non-resident withholding tax on interest paid to non-resident lenders.4 Other things being equal, such a move would have been likely to result in upward pressure on interest rates, and downward pressure on the exchange rate. Such a rebalancing of monetary conditions would probably have generated some short-term relief for the traded goods sector. The Bank took the view that these short-term benefits were unlikely to justify reversing the earlier liberalisation and reinstating controls on capital markets. Also, as with Chilean-style controls discussed earlier, there could have been potentially significant, and lasting, implications for foreign investors' confidence in New Zealand as a place to invest. That could have increased the cost of capital faced by New Zealand borrowers, not just in those periods when monetary policy was attempting to lean against strong demand, but in periods of weak demand and activity as well.
33. Another tax issue that we highlighted was the treatment of capital gains on residential investment. It had become apparent that, over the years, capital gains on an increasing range of investments had been captured within the Inland Revenue Department's definition of "assessable income". However, capital gains/losses on privately owned houses and farms, including those which are rented or leased, generally remain non-assessable for income tax. The number of avenues for "tax efficient" savings had been narrowed down over the years and, as a result, those which remain - notably private investment in real estate - may have become relatively more attractive. Combined with the tax policy developments mentioned above, the effect may have been to shift the balance of advantage against investing in the institutional sector (that is, for example, in unit trusts and superannuation funds), and towards individuals investing directly, especially in real estate. Indications from Auckland in particular suggested that there had been an upsurge in individuals buying residential real estate as a retirement savings vehicle.
34. Against this background, closer analysis of the treatment of "capital gains/losses" in the income tax system seemed warranted. Prima facie, it appears that the exclusion of capital gains/losses on real estate from taxable income may have been increasing the amplitude of real estate market cycles. If so, this effect would complicate the operation of monetary policy, since it implies a need for stronger offsetting monetary policy responses than might otherwise be required.
35. In other areas, the Reserve Bank took the initiative to commission work on the possible impact on housing and section price inflation of the way local councils were administering the Resource Management Act, and wrote to the Minister of Finance in support of allowing parallel importing. we devoted some time to analysing one other private proposal: the Interest-Linked Savings Scheme (ILSS). Under this scheme, whenever the authorities needed to raise the cost of credit, interest rates would not be raised, because to do so would simply attract capital inflows and push up the exchange rate. Instead, borrowers would be levied an adjustable ILSS surcharge, over and above the interest rate charged by the lender. Such a structure would have created very substantial incentives to avoid or evade the impost. A substantial administrative machinery would have been required to try to enforce such a scheme and substantial damage would have been caused to New Zealand's financial markets and their integration with international capital markets.
36. The discussion in this section is not intended to be comprehensive. Apart from anything else, it is not the Reserve Bank's role to do the detailed analysis of many of these superficially helpful options. It is designed, however, to illustrate the Reserve Bank's ongoing consciousness of adjustment issues. Price stability is a highly desirable feature of a modern well-functioning economy. The net benefits are increasingly widely recognised, but it is also increasingly recognised that it is in the transitions, and especially amid the sustained tightenings that are at times necessary, that the dislocations and costs of the framework become apparent.
Sterilised intervention in the foreign exchange market
37. A more orthodox tool still used by many central banks, and held in active reserve by most others, is sterilised intervention in the foreign exchange market. Sterilised intervention involves a central bank buying (selling) its own currency in the market in exchange for foreign currency. The "sterilisation" involves neutralising any impact of the intervention on the level of settlement cash in the banking system.
38. In the fifteen years since New Zealand floated its exchange rate in March 1985, the Reserve Bank has not intervened in the foreign exchange market. In New Zealand, intervention has been deliberately reserved for cases of "extreme disorder" in the foreign exchange markets: symptoms of such disorder might include the withdrawal from the market of many of the key interbank dealers, or a very marked widening of spreads between quoted buying and selling prices. The aim of such intervention would be to help restore a functioning market, recognising the potential costs and dislocations that can arise if markets stop functioning, or if it is feared that they are about to cease functioning. The goal would not be to defend a particular exchange rate, although intervening to facilitate the functioning of the market might limit some of the extreme movements in exchange rates that might otherwise have occurred. Using the "extreme disorder" criteria, the Reserve Bank came close to intervening in the early years of the float, but did not actually have to intervene; a record that we had not envisaged when the exchange rate was first floated. To support this policy, we continue to hold and actively manage around $4.5 billion of liquid foreign currency reserves.
39. Our approach to foreign exchange market intervention has been strongly influenced by our history; specifically, the history of "intervention" across the whole spectrum of policy, and the reaction against that experience in the mid-1980s. Moreover, as discussed in the paper in this series "The Evolution of Monetary Policy Implementation", in the immediate aftermath of liberalisation we had a generally very relaxed approach to financial market price volatility for some years. And through much of the 1990s our monetary policy implementation arrangements - eg exchange rate indicator ranges - tended to deliver us relatively low short-term exchange rate volatility.
40. For other countries the historical background was different. While views are divided on the effectiveness of intervention and/or the preconditions for successful intervention, and intervention strategies and techniques have come and gone over the years, few central banks have shown signs of completely abandoning the option of intervention.5 Most central banks that retaining the option of intervention appear to be more amenable to actually intervening than we have been.
41. Against that background, and the disquiet felt in New Zealand about the extent of the cyclical exchange rate fluctuations, should foreign exchange intervention have a place in the tool kit, not just in cases of "extreme disorder", but also perhaps to ease the exchange rate pressures at the margin?
42. Note that there is an important distinction here. Discussion about intervention takes for granted that a floating exchange rate is the normal state of affairs. Thus defences of fixed exchange rates, such as the Bank of England's efforts in 1992, or those of various Asian economies in the midst of the crisis, are not being reviewed here. There are arguments for and against fixed exchange rates, but we remain convinced that a floating exchange rate is appropriate for New Zealand.
43. Discussions with other central banks and reviews of the literature confirm our sense that foreign exchange intervention alone cannot change the essential character of exchange rate cycles in developed economies. Macroeconomic factors - relative interest rates, relative growth rates (both actual and expected) etc - will largely determine the character of those cycles. For example, if sterilised intervention had been used during the mid-1990s in New Zealand, the exchange rate would still have appreciated very strongly under the influence of the high New Zealand interest rates.
44. But equally, well-done interventions at times do seem to have been able to change market expectations and/or affect market behaviour in a way that may limit some of the extreme exchange rate overshoots, or perhaps take the last few per cent off some exchange rate cycles. Some of the Reserve Bank of Australia's interventions in mid-1998 look to have been good examples of the former, faced with the sort of unfavourable "market dynamics" in which one seller's behaviour is triggering more selling and so on, regardless of the underlying "fundamentals".
45. The formal statistical evidence on whether intervention moves the exchange rate is relatively equivocal. However, markets remain very wary of central bank intervention, and research suggests that interventions have typically done on the "right" side (that is, central banks have been buying their local currency when the exchange rate is below the long-run average level of the exchange rate, and selling the local currency when it was above its long-run average). Consistent with that, foreign exchange interventions appear to have been profitable, probably even after allowing for the financial risk the authorities assume in intervening. That, of course, is not a reason for intervening, but it is somewhat reassuring that whether or not the intervention is having a beneficial impact on the exchange rate, it is unlikely to be doing much harm, or costing taxpayers' money.
46. Why might intervention work? The sheer size of the intervention is unlikely to be what determines its impact. Transacting even several hundred million dollars in a market like that in, say, the Australian dollar would not normally make much difference to the exchange rate. And financing that intervention in the domestic debt markets would be unlikely to drive much of a wedge between domestic and foreign interest rates. Nonetheless, these sorts of effects should probably not be discounted altogether. The markets' willingness to be exposed to any particular market/currency, and especially that of a very heavily indebted country, is limited: particularly so when investors become very nervous or risk averse, and markets become relatively less liquid. A central bank that was willing to allow its reserves to fluctuate quite substantially over the business cycle would have a considerably longer holding or reporting horizon for its position than any normal investment institution.
47. Generally, however, foreign exchange interventions seem most likely to "work" through other channels. "Signalling" is one channel. Putting money on the line may provide markets with an added sense of a central bank's expectation (perhaps already conveyed in words in speeches or other policy statements) that a monetary policy cycle is about to turn, or that an exchange rate has seriously overshot. That suggests that intervention is more likely to work when the market is uncertain, "skittish", and thin - searching for direction or searching for a sign of which way the big players in the market are likely to jump next.
48. Does foreign exchange market intervention then have a place in the modern world? We have recognised a limited place for it all along, but have been more reticent than many other authorities in considering using the tool. However, the idea that the place for intervention is limited is probably universally accepted among countries with developed markets. Like any other crisis response tool, the less it has to be used the better. Indeed, intervention may be more effective when it is used only rarely and infrequently. In the mid 1990s, the Reserve Bank concluded that it was not then appropriate to intervene to lean against the peaks of the exchange rate cycle. After 15 years of not having intervened, the threshold for intervention remains high, and for us to intervene in circumstances other than demonstrable crisis would involve reputational issues that other countries, which have used the instrument more frequently, would not face.
1 Even more directly, in 1967 banks were simply instructed to reduce the aggregate level of approved credit limits by 10 per cent.
2 Note that banks do use administrative devices other than interest rates in making lending decisions. Loan-to-value ratio limits are common for mortgages, and in some circumstances there will be no interest rate at which a bank will be willing to lend to a particular borrower or project. However, these constraints reflect judgements about creditworthiness, and are not a mechanism for rationing a limited pool of available funds.
3 Consider a rather extreme illustration. If a ratio system was established requiring supermarkets to hold in stock, say, 10 percent of the amount of sugar they sell each week, it would have no additional impact on the volume of sugar bought by consumers, over and above the impact of that restriction on the price of sugar. Supermarkets would have to buy additional sugar and cover the holding costs of those sugar stocks. That would almost certainly raise the retail price of sugar and might deter some buyers - but only through the price mechanism.
4 The Approved Issuer Levy scheme effectively eliminates the non-resident withholding tax liability in respect of most overseas borrowings.
5 For example, in the wake of their experiences during the Asian crisis, the Bank of Canada abandoned its long-standing policy of undertaking frequent interventions to smooth short-term movements in the exchange rate.