Sterilised Interventions as an Additional Policy Instrument

  • Felix Hüfner
Part of the ZEW Economic Studies book series (ZEW, volume 23)


Theoretically, there are two possible ways for a central bank to influence the exchange rate:
  1. 1.

    indirectly via interest rates

  2. 2.

    directly through foreign exchange market interventions.



Exchange Rate Interest Rate Monetary Policy Central Bank Foreign Exchange 
These keywords were added by machine and not by the authors. This process is experimental and the keywords may be updated as the learning algorithm improves.


Unable to display preview. Download preview PDF.

Unable to display preview. Download preview PDF.


  1. 57.
    As a consequence, Kim (2002) suggests that “it is important to consider foreign exchange intervention in the studies on monetary policy and the exchange rate.” (p. 17).Google Scholar
  2. 58.
    Thus, foreign exchange intervention as a policy tool in the context of inflation targeting differs fundamentally from strategies such as fixed exchange rates.Google Scholar
  3. 59.
    While in a domestic market operation domestic assets of the central bank are used, foreign assets are used in a foreign exchange intervention. The resulting effect on the monetary base, however, is equal in both cases which is why unsterilised interventions are equivalent to domestic open market operations regarding their effects on the monetary base.Google Scholar
  4. 60.
    The question whether sterilised interventions can still be effective is analysed in detail in chapter 4.2.Google Scholar
  5. 61.
    Any US foreign exchange interventions are undertaken by the Federal Reserve Bank of New York on behalf of the Federal Reserve System and the US Treasury (see
  6. 62.
    The lagged dependent variable is included to capture other influences on NDA than sterilisation.Google Scholar
  7. 63.
    See also Goldstein (2002) as another proponent of the concept of managed floating. 64 The exchange rate is only perfectly controllable in a situation of appreciation. In a situation of continuous depreciation the level of central bank foreign exchange reserves limits the ability to control exchange rate movements (see Bofinger and Wollmershäuser, 2001:21).Google Scholar
  8. 65.
    In their model, the MCI consists of the real interest rate and the real exchange rate. However, since they assume sticky prices in the short run, both variables can be controlled directly with their nominal counterparts.Google Scholar
  9. 66.
    For the Canadian MCI the weights of interest rates to exchange rates are 3:1 and for the New Zealand MCI the weight of interest rates to exchange rates is 2:1. For the New Zealand MCI this means that a 2% rise in the exchange rate has the same effect on future inflation as a 100 basis point (one percentage point) increase in short-term interest rates (accordingly, the relative influence of the exchange rate versus the interest rate is smaller in Canada with its 3:1 weighting). See Bernanke et al., 1999.Google Scholar
  10. 67.
    This inability of central bank to stop sudden and massive capital outflows is exemplified in the event of currency crises: a fixed exchange rate can only be defended as long as the central bank has enough foreign currency reserves.Google Scholar
  11. 68.
    See for the following Bofinger and Wollmershäuser (2001).Google Scholar
  12. 69.
    See BIS, 2001, Table 4.Google Scholar
  13. 70.
    There are only a number of central banks which provide official intervention data as a time series for researchers (to our knowledge these are the Bundesbank, the Swiss National Bank, the US Federal Reserve, the Reserve Bank of Australia and a number of smaller emerging market countries). Some central banks provide on their websites information about interventions on the day they occur but do not publish them in time series format (like the Bank of Canada or the Sveriges Rikbank). Usually, also the volume of intervention is kept secret.Google Scholar
  14. 71.
    This paragraph draws on Taylor (1995) and Spolander (1999).Google Scholar
  15. 72.
    If purchasing power parity holds continuously, the real exchange rate is a constant.Google Scholar
  16. 73.
    Consequently, the uncovered interest parity condition does not prevail in portfolio balance models (see Isard, 1995: 107). It does, however, prevail in monetary models since perfect substirutability in these models does not require that investors be given a risk premium for holding a particular currency.Google Scholar
  17. 74.
    One study that finds significant effects for the portfolio balance channel in interventions of the early 1980s is Dominguez and Frankel (1993b).Google Scholar
  18. 75.
    The discrepancy between actual and correctly reported intervention might also stem from the fact that the Fed conducted customer transactions in the markets that it not published as official interventions (see the discussion in the chapter on foreign exchange reserves).Google Scholar
  19. 76.
    A Bloomberg news report cites a Bank of Canada money trader as saying “The Bank of Canada won’t announce when it’s buying or selling currencies to help the dollar […]. ‘There will not be an official statement ‘about sales or purchases, said Christian Dupont, a currency trader at the central bank. We will not publish anything and will not go ahead without the market asking‘.“ (Bloomberg News, March 25, 1999, 4.14 PM, Bank of Canada to confirm currency market activities).Google Scholar
  20. 77.
    The portfolio balance channel also gives no justification for secret intervention operations.Google Scholar
  21. 78.
    Dominguez and Frankel (1993 a) argue similarly: “How can intervention signal future monetary policy intentions if the market does not hear the signal?” (p. 71).Google Scholar
  22. 79.
    Djoudad et al. (2000) provide evidence of the existence of chartist traders in the foreign exchange markets of Australia, Canada and New Zealand. Allen and Taylor (1990) find that the importance of chart analysis declines over the forecast horizon. While 90% of their survey respondents use chart analysis for short horizon currency forecasting (intraday up to one week), fundamental analysis becomes more important at forecast horizons of one year or longer.Google Scholar
  23. 80.
    Frankel and Froot (1990) state: “It has long been remarked that if there exist traders who tend to forecast by extrapolating recent trends (i.e., who have ‘bandwagon expectations’), then their actions can exacerbate swings in the exchange rate.” (p. 183).Google Scholar
  24. 81.
    Nonetheless, the central bank itself might be tempted to move the exchange away from its fundamental value which, however, is not an issue in the literature.Google Scholar
  25. 82.
    Saacke (2002) provides further insight into the relationship between the role of interventions in breaking existing exchange rate trends. He finds that exchange rates tend to move counter to central bank’s intentions shortly after interventions but that this effect reverses in the long run. This might be an explanation for the finding of many studies that cast doubt on the effectiveness of interventions by analysing only a short period around intervention operations. As is illustrated in Figure 9 it can take some time before the trend is reversed; all interventions before the actual trend reversal would be interpreted as ineffective in the short run. In the long run, however, the operations proved to be rather successful. Neely (2002) finds evidence for some countries that high returns from trading rules (technical trading) precede intervention operations which supports the hypothesis that interventions are intended to break existing trend lines.Google Scholar
  26. 83.
    This serves also as an explanation for the observation that central banks often decide to enter markets when volume is low, for example late in the Asian trading session or before weekends, as the price impact of the intervention is larger and the likelihood of breaking an existing trendline is higher.Google Scholar
  27. 84.
    This view is supported by the empirical analysis of Ramaswamy and Samiei (2000) who find that “interventions have succeeded on a number of occasions during 1995-99 in changing the path of the yen-dollar rate in the desired direction.” (p. 23).Google Scholar
  28. 85.
    This result is common knowledge since the early 1980s following the seminal paper by Meese and Rogoff (1983).Google Scholar
  29. 86.
    Peiers (1997), for example, finds evidence that intervention is associated with asymmetric information flows among foreign exchange traders and that “certain commercial bank dealers have privileged information regarding central bank activities.“ (p. 1607). Her results suggest that Deutsche Bank has been a price leader up to 60 minutes prior to Reuters news reports of Bundesbank intervention activity. Dominguez (1999) reports similar results: she finds significant exchange rate returns up to one hour prior to Reuters news reports of intervention operations by the US Fed, the Bank of Japan and the Bundesbank.Google Scholar
  30. 87.
    In the survey by Cheung and Wong (2000) 56% of respondents state that large market players have a competitive advantage that stems from a larger customer base as well as better information about the market as a whole. The former gives them better knowledge about customer order while the latter is due to a better market network that allows large traders to gain knowledge about the order flow of other banks.Google Scholar
  31. 88.
    In 2001 the US$/Euro currency pair was by far the most traded with a share of 30%, followed by US$/Yen (20%) and US$/GBP (11%) (BIS, 2001).Google Scholar
  32. 89.
    Dominguez and Frankel (1993a) report for the 1980s an average intervention volume of 200 million US$ per day. However, volume can be assumed to have increased since then. Official intervention data of the Bank of Japan, for example, show an average volume of 1.2 billion US$ per day for the period 1991-2001 with the largest intervention occurring on April 10, 1998 when the Bank of Japan sold an equivalent of 20 billion US$ in support of the Yen.Google Scholar
  33. 90.
    Proponents of the microstructure approach also use this argument against the idea of Tobin taxes on foreign exchange transactions. Since according to inventory control the high volume simply represents risk management of foreign exchange dealers rather than excessive speculation, such a tax would seriously impede this risk management (see Lyons, 2001b: 13).Google Scholar
  34. 91.
    Surely, one could take a different approach and argue that the existing practice of central banks lacks a theoretical fundament and therefore is doomed to failure. However, central bank intervention occurred frequently over the past three decades. We doubt that agents in central banks would use this tool without proper theoretical justification and thus proceed to try to understand the reasoning behind these operations. 92 As mentioned above, it might be asked why the central bank sterilises the intervention at all if the intention is to signal future monetary policy. This question, however, does not occur in discussions of the signalling channel.Google Scholar
  35. 93.
    A further argument for secret interventions can be derived from a statement in Mishkin (2000): “Responding too heavily and too frequently to movements in a ‘flexible’ exchange rate runs the risk of transforming the exchange rate into a nominal anchor for monetary policy that takes precedence over the inflation target, at least in the eyes of the public.” (p. 109). In order to let the public believe that the inflation target is the primary goal, central banks may chose to keep interventions secret.Google Scholar
  36. 94.
    A different but important question is if these theoretical models do justify interventions at all. In order to make sense of interventions the underlying exchange rate model has to allow for the possibility of exchange rates deviating from their fundamentally justified range. This is not the case in monetary and portfolio-balance models, but is an essential part of the noise trader approach. In the signalling channel interventions are necessary for signalling future monetary policy. Since the microstructure approach is no pure exchange rate model it does not tackle the issue whether interventions are necessary at all (though Lyons (2001b: 230) argues that the central bank knows that the current exchange rate is inconsistent with fundamentals whenever there are significant imbalances between buyer-and seller-initiated orders).Google Scholar
  37. 95.
    Indeed, the minutes of a recent monetary policy meeting of the Swedish Riksbank include a discussion of the noise-trader and microstructure approach with regard to their relevance for central bank interventions (see Riksbank, 2001b). Riksbank (2002), however, attributes an important role for the signal channel.Google Scholar

Copyright information

© Springer-Verlag Berlin Heidelberg 2004

Authors and Affiliations

  • Felix Hüfner
    • 1
  1. 1.Centre for European Economic Research (ZEW)MannheimGermany

Personalised recommendations