Why Do Capital Flows Cause Problems?

1. Macro Problems

These can be grouped into three:
• Inflation pressures
• Exchange rate problems: appreciation of the exchange rate (uncomfortable for the traded sector) combined with a volatile/fragile exchange rate
• Loss of control over, or constraints on the free use of, monetary policy

(a) Inflation

We argued above that capital flow adds as much to supply as it does to demand, so it should not, in itself, be inflationary. It might, however, be inflationary in other ways. If it makes monetary policy less effective or constrains the use of monetary policy, this might be relevant, but we will argue below that competent monetary policy can avoid this threat.

It does, however, leave open the possibility (in fact the likelihood) that the inflows will put pressure on asset prices. There are two possible inflationary channels. Although capital flow adds to supply, the inflow may be directed mainly at the purchase of existing assets (physical assets and equities might be thought of as non-traded goods), so the demand pressure is not offset much by the extra supply of cheap foreign goods and services. This asset price inflation is accommodated by the stance of domestic monetary policy, which targets CPI prices rather than asset prices. Central banks remain uncomfortable with this asset price inflation as it is distortionary while underway, exacerbates the cycle and is disruptive when the asset bubble eventually bursts. This was certainly the case in a number of East Asian countries in the years leading up to the Crisis. This does present a dilemma for policy: the authorities can raise interest rates, but they refrain from this (or do it only in moderation) partly out of the belief that their actions will be to some extent frustrated by extra capital inflows, but mainly because they do not believe they can effectively control asset prices and do not want to be blamed for pricking the bubble when the asset prices eventually fall. They have, moreover, judged themselves to be unable to do more than, at most, lean against the wind, ready to pick up the pieces when the asset bubble bursts. This is unsatisfactory, but represents the imperfect current “state of the art.”

If the inflow does bid up asset prices, there is a second possible effect. It increases the general investment demand via Tobin’s “q”: existing assets are now more expensive than the cost of reproducing them, so investment is boosted. Higher equity prices cheapen funding costs, so thus encourage more investment.

(b) What Is the Analytical Model for the Exchange Rate?

Cyclical pressures on the exchange rate (as it acts as an important channel for monetary policy) combine with the structural influences to produce an exchange rate which has a strong tendency towards appreciation, and has no clear anchor in the “fundamentals.”

It is possible to explain the cyclical path of the exchange rate in terms of the Dornbusch (1976) over-shooting model (although it is hard to identify this process in the real world), just as some cyclical movements may be explicable in terms of the world commodity-price cycle (see Gruen and Kortian, 1996). Economic analysis, however, has little to say about the path of the exchange rate during the decades-long journey to the technological frontier. It might be possible to envisage the exchange rate being on a steady trajectory towards the longterm equilibrium, when the economy has reached the technological frontier. But at any point on this path, this exchange rate will be too low for portfolio equilibrium, as the investors face the prospect of higher interest returns and exchange rate appreciation. Suppose the rate appreciates sufficiently to fully anticipate the end-point of the structural appreciation (some decades ahead): the investor still has the advantage of the higher interest rate in the meantime. So nothing short of a once-off appreciation beyond this long-term equilibrium, followed by a steady depreciation (rather like a very drawn out version of the Dornbusch over-shooting process) would maintain portfolio equilibrium.

We observe that this inflow is not equilibrated by price arbitrage: the foreign and domestic interest rates do not merge together over time. So there must be other forces at work constraining the inflow. One common approach is to explain the enduring interest differential as a “risk premium.” This might mechanically satisfy some portfolio balance constraint, but is analytically unhelpful unless some explanation can be offered for the risk premium and how it changes over time.

Is there a structural analogue of the cyclical Dornbusch overshooting mechanism? If the exchange rate appreciates and remains above its longer-term equilibrium until some random shock creates the risk of a short-term fall, the prospect of even a small fall in the near future outweighs the interest differential. This would have to be a very random, tenuous and unstructured equilibrium path because a longer-term investor would not be deterred by this short-term depreciation risk or high-frequency volatility. Investors with a short-term horizon, however, might want to cut their exposure. We might expect to see not only swings in the exchange rate of the recipient country, but in the capital-supplying country as well. This fits well with the experience of Japan during the yen-carry period: the substantial current account surplus, together with an undervalued exchange rate (a real rate which is lower than in the 1990s) punctuated by a sudden sharp appreciation whenever the outflows are in question (October 1998, August 2007) with very large swings (in the range of 80-150 yen/dollar).

This fits with the idea of “sudden stops.” Some simply call this a “time-varying risk premium” and leave it at that. The more honest approach is that taken by Krugman (2006), who calls this a Wyle E. Coyote process: “a moment when investors realize that the dollar’s value doesn’t make sense and that value plunges.”⁵ This puts the sophistication of the analysis on the right level: that of a cartoon. The “search for yield” lasts as long as asset prices are rising and the boom is strong. The most plausible explanation of the Asian contagion in 1997 is Morris Goldstein’s “wake up call”: nothing more substantive than a reminder that there was an issue. More often than not, the trigger for outflow is an external policy event rather than a domestic one (see Feruci et al., 2004).⁶ Perhaps an insight can be gained by remembering that the foreign investors usually know very little about the specifics of their investment or even the country they have invested in. The arrival of a small amount of new information can add hugely to their stock of knowledge, and lead to an abrupt change of view.

The markets themselves encompass self-exacerbating processes. They use similar risk models, which signal the same decision-point for all investors. Credit rating agencies set their evaluations by looking in the rear vision mirror, and when they downgrade in response to bad news, investors (often driven by rating-specific mandates) are forced to sell. Herding (“if others are getting out, what do they know that I don’t?”) or “correlated errors” cause the investors to cut their investments at the same time, often into “crowded markets” where others are doing the same, with a large impact on prices. A fall in the exchange rate is supposed to create the expectation of a subsequent rise (“mean reversion”), but when the exchange rate is unanchored, it can fall greatly without encouraging new inflows (as seen during the Asian crisis). Eventually, however, the fall ends and the underlying interest differential reasserts itself, setting off a new exchange rate cycle.⁷

This creates the possibility–in fact the likelihood–of broad swings in the value of the currency, perhaps following the periodicity of the business cycle. This is not an issue of short-term volatility of the exchange rate (the usual subject of economic analysis and market risk-analysis), but of sustained departures from the equilibrium exchange rate: misalignment rather than volatility.

It is hardly surprising that policymakers find this world–an overly-appreciated exchange rate with a tendency to sudden gyrations–uncomfortable and unattractive. In most cases in East Asia since the crisis, with flexible exchange rates in place, the policy concern has not been that capital flows threaten price stability, but rather that the inflows set in train this appreciation/instability of the exchange rate.

The appreciated exchange rate undermines international competitiveness, at the cost of slower growth in the tradable sector, often the most dynamic sector of the economy (for argument in favor of under-valued rates and further references see Rodrik, 2007).⁸

The greater the appreciation, the larger the fall when the reassessment comes. With the addition of some overshooting in the opposite direction, spill-over into inflation and selfreinforcing capital flight (examined in the next section), the stage is set for a crisis.

With the exchange rate subject to this sort of random influence, policymakers face the difficult task of distinguishing between this randomness and the on-going and continuous changes in the equilibrium, with the danger that they may try to resist the latter as well as the former. However uncomfortable it may be, the authorities in the emerging countries have to accept the need for some appreciation. A capital inflow should put upward pressure on the exchange rate, because this is the mechanism through which the real counterpart of the financial capital inflow–the transfer of resources–takes place: the appreciation encourages imports and discourages exports.⁹ This is true whether the capital flow is long-term structural, or cyclical. In both cases the movement of the exchange rate is part and parcel of the adjustment process, and policy should not resist it. Its unwelcome nature is, however, understandable: even if the authorities acknowledge that this sequence–with appreciated exchange rate and current account deficit–is the necessary channel for the capital inflow to operate, they no doubt recall that both these same elements–appreciation and CAD–were identified as being central causes of the Asian crisis and often blamed for the problems (see, for example, Feldstein, 2000). Misguided though such criticism might have been, it was important in undermining confidence. Policymakers are understandably reluctant to leave themselves and their countries open to a repeat performance.

We might note in passing how the Impossible Trinity led to a focus on the wrong issue after the crisis. The exchange rate debate focused on the exchange rate regime: specifically on the need for “corner solutions” (the rate should either be immutably fixed or a pure free float). The middle ground of managed rates was out of bounds. Over time, opinion has softened and fuzzed (see Fischer, 2002) and now focuses, more narrowly and sensibly, on the dangers of a fixed-but-changeable peg. In the meantime, however, attention was distracted from the possibility–indeed the likelihood–that at times the unanchored exchange rate will be significantly away from its equilibrium value and for long enough to do damage. In the fixed/free-float dichotomy, policymakers have no need to think about some notion of the “right” level of the exchange rate. But if the middle ground of partially managed rates turns out to be the practical reality, then policymakers need a framework in which the value of the exchange rate has some place in their policy consideration.

Of course it is not easy to operationalize such a framework, and many will see this as a distraction from the single-objective approach to monetary policy. However, for countries that are not yet ready to let their shallow and immature foreign exchange market handle the price discovery (i.e., they retain a “fear of floating”: see Calvo and Reinhart, 2002), there is a vital need to have some fairly specific working notion of what is the “right” exchange rate (if only in terms of a range), and how this might change the cycle structurally over the medium term. They also need some notion of how to reconcile the possibly conflicting signals which the foreign exchange market may be giving to their price stability objectives.

We will return to this issue below, when we discuss policy measures. For the moment it is sufficient to observe that exchange rates in emerging countries are not well anchored by widely-accepted stable views about the “fundamentals” or a long track-record which can establish the parameters of a mean-reverting process, and while memories of the huge movements during the 1997 crisis remain, exchange rates will be vulnerable not just to short-term volatility, but to sustained misalignment.

(c) Loss of Monetary Control

The “Impossible Trinity” was a warning that countries could not simultaneously have open capital markets, a fixed rate, and independent monetary policy. A floating rate would free up monetary policy, but interest rates and the exchange rate would be governed by uncovered interest parity (UIP): the interest differential had to remain equal to the expected change in the exchange rate. Thus, if interest rates were altered for domestic purposes, the exchange rate would move in response, mapping out the overshooting path prescribed in Dornbusch (1975). The corollary was that the exchange rate could not be influenced without changing the domestic interest rate.

Meanwhile, in the real world, major countries with a high level of integration into world financial markets experienced decade-long periods with substantially different interest rates (see Figure 2 [ PDF 20.6KB | 1 pages ]). US and Japan had an average differential of more than 300 basis points–with the US rate being three times the Japanese rate–for most of the past fifteen years. It is true that these countries had floating rates, but there was no sign of the Dornbusch portfolio equilibrating process at work. It would have required a once-off step “overshooting” depreciation of the yen followed by a steady appreciation. Instead, the cross rate has fluctuated between 90 and 150, with three wide cycles over this period.

More relevant to this paper, the countries of East Asia have by and large been able to set policy interest rates where they wanted them, both before the crisis (when the exchange rate regimes were semi-fixed, and interest rates were routinely higher than world rates) and since (when the regimes are usually classified as a “managed float”): see Figure 3 [ PDF 114.8KB | 1 pages ]. Despite very large capital inflows, interest rates have not equalized, even for countries with relatively small financial sectors. In Thailand, with a fixed rate, foreign savings equivalent to 9 percent of GDP in the single year of 1996 were insufficient to bring interest rates into line with foreign rates.

Why is reality so far from the message of the Impossible Trinity? First, UIP assumes perfect substitutability between domestic and foreign assets, even when these are denominated in different currencies. This would require much more than the absence of capital controls: it would need well-developed institutional connections, full information (often about countries with very different systems and stage of development), similar tax and legal regimes, similar risk appetites and, above all, a very clear view about the future path of exchange rates.^{10 , 11}

Secondly, UIP posits a very direct connection between capital inflows and looser monetary policy, because it envisages the credit multiplier process as being the basis of monetary policy, whereby a rise in foreign exchange reserves adds to base money, which is multiplied up automatically into credit growth. With financial deregulation, this model is no longer relevant, and it is now feasible, within broad limits, for the authorities to maintain the policy interest rate in the face of capital inflows.¹²

With that background, we can now ask if capital flows cause any loss of control over money supply. First, we can test this against the pre-deregulation common target–base money. This has special resonance for the Impossible Trinity, because this was the channel through which attempts at policy differentials would be frustrated. Table 1 [ PDF 15.8KB | 1 pages ] shows M1 and M3 growth compared with growth in foreign exchange reserves. It is difficult to see any connection.

Figure 4 [ PDF 77.2KB | 1 pages ] presents another view of this issue. With the possible exception of India, there does not appear to be any close link between additions to net foreign assets and base money.

Why is this linkage so weak? First, the process of sterilization seems to have been quite effective. In practice it is relatively easy for central banks to sterilize excess base money in a deregulated system, as banks have no alternative use for it if they are already supplying all the loans that are demanded at the going policy-based interest rate. In any case, where the interest rate is the policy instrument, there can be a great deal of slippage between base money and credit (which is the money variable that impinges directly on economic activity). If the authorities have set the interest rate structure, this will determine the rate of credit expansion, and excess base money may not have much effect on credit growth: it remains as unintended excess reserves in the banks’ balance sheets (c.f., Japan in 2001-2004 and Indonesia in 2005-6).

We can also examine whether the authorities are able to maintain their policy interest rates in the face of a large build-up in foreign exchange reserves. Ho and McCauley (2008) conclude that: “Central banks with explicit short-term interest rate operating targets or official rate corridors (for example, in India, Indonesia, Republic of Korea, Malaysia, the Philippines and Thailand) were able to manage money market liquidity such that the relevant interest rates did not fall and stay below their announced targets, notwithstanding bouts of foreign exchange purchases.” ¹³

It is possible that the authorities would have preferred higher interest rates and trimmed their setting in the hope of discouraging some of the excessive capital inflow. But if they did trim their policy instrument, it does not seem to have stopped them from achieving their final objective–low inflation. So far this century, despite very large capital inflows, inflation has byand- large been contained (although the People’s Republic of China [PRC] may represent an unfinished story). Ho and McCauley (2008) conclude:

“All in all, Asia during the period under consideration did not provide evidence for the well-known argument that large-scale reserve accumulation would be inflationary. The top reserve accumulators, be it in absolute terms (PRC and Japan) or in relative to GDP terms (Singapore; Malaysia; Taipei,China; and PRC), did not experience notably larger rises in inflation over the period 2002-2006 compared to economies that accumulated little reserves.

More strikingly, there is in fact an inverse relationship between reserve accumulation and average inflation performance in Asia over the same period. The top reserve accumulators all had relatively low inflation or even deflation. In contrast, two economies that saw the least reserve accumulation (Indonesia and the Philippines), given currency weakness through 2005, were the ones that over-shot inflation targets and experienced the highest inflation in the region. This inverse relationship is even more evident if one juxtaposes the inflation rate in 2001 (i.e., the initial condition) with the subsequent degree of reserve accumulation.” (p. 11)

Not everyone shares this assessment. The BIS 2007 Annual Report, using a wider range of countries, claims to see some relationships between, on the one hand, growth of foreign exchange reserves and, on the other, base money, credit and inflation (see Figure 5 [ PDF 170KB | 1 pages ]). It has to be said, however, that these look to be pretty tenuous relationships, relying on a couple of outlier countries for their visual impact, with little general explanatory power.

The IMF, stuck as usual in a decades-old paradigm, still wants to test the Impossible Trinity in terms of the relationship between base money and credit growth (see IMF World Economic Outlook, October 2007).

2. Flighty Volatile Capital: Sudden Stops

If a flexible exchange rate is not well anchored by expectations and a well-established history of mean reversion around some longer-term trend, “sudden-stop” capital reversals are a constant danger. As capital leaves in response to a disturbance or change in confidence, it drives down the exchange rate, causing a vicious cycle as more capital leaves in response to this fall.

When these investors flee, they are not easily replaced by other foreigners. These investors in emerging countries are on the frontiers of fund-management, not the mainstream. Few other foreigners can be persuaded to invest by a modest fall in the exchange rate because the exchange rate is not well anchored and there is no general perception of what the “right” rate is.¹⁴ Capital inflow in emerging economies is binary: it is either on or off.

Such sudden outflows require a huge and painful adjustment process. When the capital flow is inelastic in response to a lower exchange rate, the adjustment has to take place largely in terms of income falls (reduced absorption) which, through reduced imports, are the only path by which the current account can be quickly brought into equilibrium with the now-reduced foreign funding. The exchange rate cannot produce a quick response by “switching,” so the equilibrium has to be achieved by painful “adjusting.”

To illustrate the point, let us compare Australia and Thailand during the Asian crisis. The fall in the Australian dollar was not, of course, as great as in Thailand, but it was nevertheless very substantial–close to 30 percent. The relationship between this exchange rate fall and capital flows was, however, quite different, for reasons we will explore in a moment. But first, let us look at the data. Figure 6 [ PDF 18.6KB | 1 pages ] shows the huge turnaround in capital flows in Thailand (amounting to well over 20 percent of GDP, from an inflow of nearly 9 percent in 1996 to an outflow of 13.6 percent in 1998. This contrasts with Australia (shown in Figure 7 [ PDF 18.6KB | 1 pages ]). At least in this annual data, there is no sign of any reversal of capital at all in the case of Australia, despite the significant fall in the exchange rate: the inflow was actually larger in 1998 than the previous average.¹⁵ Thus for Australia the fall in the exchange rate was a threat to inflation (which in the event came to be seen as a tolerable threat, as the pass-through was much slower/smaller than had previously been thought), but not to capital flows. Relaxed about the threat to price stability, the Australian central bank was prepared to let the exchange rate fall without raising interest rates in its defense.¹⁶ The result was that the real economy was largely unaffected (if anything, it was stimulated by the lower exchange rate). The Thai authorities, on the other hand, were forced to raise interest rates in an economy already put in free-fall by the need to trim the current account to the available (hugely reduced) foreign funding. Clearly there is a different relationship between exchange rate weakness and capital flows.

Others countries provide similar comparisons. Cabellero et al. (2004) argue that the different behavior of Chile, compared with Australia, was not caused by different views on the inflation danger (the pass-though in both countries is similar), but rather was aimed at pre-empting capital outflow which would be much more likely to happen in Chile (fewer opportunities to diversify risk through derivatives), and do more damage when it did (commercial balance sheets are quite exposed to exchange-rate risk).^{17 , 18} Hausmann (1999) compares Mexico and Australia: in Mexico’s case it is not clear whether the interest rate increase was a response to the inflation threat or designed to encourage capital to remain, but the capital flow behavior in the two cases is clearly quite different.

This difference between Mexico, Chile, and Thailand, on the one hand, and Australia (and similar countries) on the other is the central policy issue: what is it that makes investors prepared to hold their positions^{19 , 20} in the case of Australia, but not with the other countries? It is not that all investors have somehow covered their currency exposures in the case of Australia but not in Thailand: Australia has had a long history of current account deficits and this cumulated inflow means that someone (in Australia’s case, foreigners) is holding a very substantial currency exposure, which must give them concern as the exchange rate falls. Caballero et al. identify the difference as “country trust,” as distinct from “currency trust”:

Currency-trust describes the degree of confidence foreign investors have in holding assets denominated in the currency of the particular country. It indicates that investors believe currency movements will not be used to expropriate their investment but also that the central bank has enough control over the currency that random shocks are unlikely to lead to perverse exchange rate dynamics. In this way currency-trust is seen to be related to the concept of inflation credibility. Country-trust describes the degree of confidence foreign investors have more generally in the country, incorporating the commitment of the country to repay debts, corporate governance, the financial system and the economic stability of the country. Importantly, country-trust means that there is no need for highly specialized knowledge to invest in the country (for example about government and institutions). (Caballero et al., 2004, p. 2)

Others would describe this differently, with different characteristics. They might talk in terms of institutions and the environment of law and governance. Others would emphasize that the central issue is the disparity in size between the financial markets of the emerging countries and those of the mature countries which are the source of the disruptive flows (see Volcker, 1999; Richards, 2002; and Runchana, 2007). Still others will argue that at each stage of the exchange rate fall in Australia, foreign investors thought that the rate had fallen enough and there was no expectation of a further fall.²¹

3. Financial System Stability

It is hard to insulate the financial sector from these problems–“twin crises” are the norm (see Kaminsky and Reinhart, 1999). To start with, the capital inflow often comes through the financial sector, intermediated by banks or finance companies. Even where it does not, the banks’ customers–the companies whose balance sheets have been seriously damaged by the outflow–have borrowed from the domestic banks and they now routinely default not just on their foreign debts, but their domestic debts as well.

Both the vulnerability to capital reversals and the fragility of the financial sector reflect institutional and reputation weaknesses. The policy problem is that reputation and institutions²² cannot be built quickly or easily. The prescription is simply unattainable in the short or even medium term. While embarking on this journey towards deep and resilient financial markets, policymakers have to put in place strategies to cope with the journey. We turn, now, to that issue.

Download this Discussion Paper [ PDF 557KB| 36 pages ].

Post a Comment

We welcome your feedback on this publication. Post a comment. ADBI is not obliged to acknowledge or publish comments and may abridge or edit them before web posting.

Comment(s)

There are [0] comment(s) for this entry. Post a comment.