تاثیر مداخله بانک مرکزی بر ناهمگونی پیش بینی نرخ ارز

We investigate the impact of official foreign exchange intervention on forecast heterogeneity, on the basis of a sample of forecasts made by a large number of commercial banks over two distinct periods, for the DEM (or EUR) and the JPY against the USD. We show that heterogeneity increases as a result of foreign exchange intervention. In the case of the DEM–EUR/USD market this increase is due to unexpected intervention, while for the JPY/USD it is due to expected intervention. Our results also emphasise the role of rumours, especially for the JPY/USD. In sum, official interventions are shown to move market opinions, albeit differently across the two markets. J. Japanese Int. Economies21 (1) (2007) 38–63.

Since the publication of the Jurgensen report (1983), suggesting that official foreign exchange market intervention had little impact on exchange rates, the effectiveness of such interventions has been controversial. In trying to gauge the effectiveness of intervention more recent research has sought to disentangle the portfolio channel from the signalling channel.1 For example, using survey data on exchange-rate expectations, Dominguez and Frankel (1993) demonstrated that official interventions do have an impact both on the risk premium (portfolio channel) and on exchange-rate expectations (signalling channel). They also highlighted the importance of using interventions that are publicly advertised, as opposed to “secret” interventions, i.e., interventions not reported in the press or by wire services. Finally, Dominguez and Frankel showed that concerted interventions by several central banks were more likely to be successful. Using different empirical techniques, and sometimes a looser definition of successful intervention, Humpage, 1999, Ramaswami and Samei, 2000 and Fatum, 2002 also found official interventions to be successful, especially when co-ordinated. However other authors have found little empirical evidence of official interventions moving the exchange rate in the desired direction: when significant, interventions were often found to push the exchange rate in the “wrong” directions; that is, a dollar purchase leads to a depreciation of the dollar (Edison, 1993, Baillie and Osterberg, 1997b and Beine et al., 2002). The recent release of official data concerning Japanese interventions has moved the consensus back towards some effectiveness of interventions Ito, 2002, Dominguez, 2003, Fatum and Hutchinson, 2003, Chaboud and Humpage, 2003, Castrén, 2004 and Morel and Teiletche, 2004. This result, however, remains fragile since it is mainly based on a particular episode (Japanese large interventions under Mr. Sakakibara directorship of the International Finance Bureau of the Ministry of Finance) and/or on a specific empirical technique (the event study methodology), and Galati and Melick (2002) and Galati et al. (2004) do not provide support for such effectiveness using standard time series techniques. The empirical literature has also devoted much attention to the impact of official interventions on exchange-rate volatility. This literature generally concludes that interventions tend to raise exchange-rate volatility, be this volatility measured as the realized variance (Beine et al., 2004), by conditional volatility Baillie and Osterberg, 1997a, Baillie and Osterberg, 1997b, Dominguez, 1998, Beine et al., 2002 and Beine, 2004 or by implied expected volatility as recovered from option prices (Bonser-Neal and Tanner, 1996, Galati and Melick, 1999 and Frenkel et al., 2003).2 Traders seem to be aware of this phenomenon since, according to Cheung and Chinn's (2001) survey, 61% of US traders believe that official interventions raise volatility. As a matter of fact, except in the post-Louvre period (1987–1990), reducing exchange-rate volatility does not seem to have been the main motivation of interventions which seems rather to be explained by the wish to lean against the wind or to correct misalignments (Galati and Melick, 1999, Neely, 2001 and Ito, 2002). On the theoretical side, recent research has focused on two stylised facts of official interventions that do not accord with the so-called signalling channel, namely: (i) the existence of secret interventions, i.e., interventions that are not publicly advertised (see Neely, 2001 and Beine and Lecourt, 2004, for some recent evidence); and (ii) the low predictive power of interventions for future monetary policy (Dominguez and Frankel, 1993). Theoretical justification for secret interventions has been provided by models based on information asymmetry between central banks and traders (see, for example, the survey by Baillie et al., 2000). Such asymmetry yields an additional channel which encompasses the coordination channel suggested by Sarno and Taylor (2001) and both the noise trading and market microstructure channels listed by Frenkel et al. (2001). This ‘composite’ channel has the additional attractive feature of not being at odds with sterilization in that, in contrast to the signalling channel, it does not imply any relationship between interventions and monetary policy at any horizon. Examples of such an approach are provided by Bhattacharya and Weller (1997) and Vitale (1999). These authors assume that the central bank has insider information which can be revealed to the market through public or secret interventions. The case for secret interventions comes from the fact that the objective of the central bank in terms of exchange rate level may differ from the one consistent with the fundamentals. When interventions are carried out secretly, the progressive spreading of the information across the market opens up the range of market expectations. Based on different premises, Hung (1997) also shows that, when the central bank wants to reverse a strong exchange-rate trend, a secret, volatility-enhancing strategy may be more efficient than a volatility-reducing one. The idea is to use secret interventions with the wind when the exchange-rate temporarily moves in the “correct” direction in order to make noise traders think there is a market momentum in this direction. This result is interesting in that it relates intervention efficiency to a rise, not a fall, in exchange-rate volatility.3 This is consistent with de Grauwe and Grimaldi (2003) who suggest that transparent and predictable interventions could stabilize the exchange rate by reducing the share of noise-traders. In contrast to these views, Morris and Shin (2002) show that some imprecision in the disclosure of public information (here, through the use of secret information) could, under specific circumstances, have a stabilizing impact in that in encourages private agents to still use their private information.4 This argument relies on the idea that public authorities may not themselves know the true level of the fundamental level whereas the markets have independent sources of information.5 Going a step further, Popper and Montgomery (2001) assume that the central bank does not produce its own information, but rather picks up pieces of information from some specific informed traders. Official interventions are then viewed as a way to aggregate the individual information and to transmit it to this group of informed traders. Popper and Montgomery then show that the variance of the exchange rate is lower the larger the proportion of informed traders. Hence secret information could have relatively more impact on volatility. In general, the rise of exchange-rate volatility following an official intervention would be the by-product of the coordination channel: official interventions could be designed to break a non-desirable trend by convincing some traders to change their mind. This interpretation is supported by some empirical microstructural evidence: using Bank of Canada intervention data D'Souza (2001) finds that the effectiveness of central bank interventions is partly determined by market wide order flows which are generated subsequent to the intervention. Such flows are caused by dealers, who find that central bank interventions provide useful information about future fundamentals. It seems likely that this kind of intervention effect will increase the post intervention distribution of expectations. Indeed, Naranjo and Nimalendran (2000) have argued that dealers increase spreads at the time of interventions to protect themselves from greater informational asymmetry, while Dominguez (1999) shows that official interventions are a source of information asymmetry in that some dealers receive early information on central bank intervention relative to other traders. Information asymmetry is more generally supported by the finding that technical rules are profitable on the foreign exchange market (see, for instance, Neely and Weller, 2001) and by the heterogeneity of exchange rate expectations across individuals MacDonald and Marsh, 1996 and Bénassy-Quéré et al., 1999. Evans (2002) shows that market heterogeneity is a major source of exchange-rate volatility on the DEM/USD market. However the impact of official interventions on the heterogeneity of private expectations has not been tested directly so far. This is the topic of the present paper. In this paper we seek to explore some of the above-noted issues and, specifically, we study whether official interventions from the Federal Reserve, the Bank of Japan, the Bundesbank and the European Central Bank (ECB) had an impact on forecast heterogeneity. We interpret our analysis as a test of the so-called coordination, or microstructural, channel of central bank interventions, namely that official intervention, or, more specifically, secret interventions, can result in an opening up of the range of market expectations. Our analysis involves the Deutschemark, the euro and the yen against the US dollar. For our measure of heterogeneity, we use the cross-section coefficient of variation derived from Consensus Forecasts monthly survey data over the periods 1992–1994 and 1996–2001.6 We subsequently use wire reports to disentangle those interventions which had been anticipated from those which took the markets by surprise. Our main results may be summarised in the following way. We find that official interventions had a positive impact on forecast heterogeneity, although in a different way across markets. For the Deutschemark and euro against the US dollar, we find that only news of interventions had an impact, whereas for the yen against the US dollar, forecast heterogeneity is moved by false rumours (i.e., rumours of interventions that were not followed by actual interventions) as well as by interventions which had been expected. Since the latter, by construction, are not secret interventions, the results for the JPY/USD seem to only weakly support the microstructure literature suggesting that especially secret interventions could lead to more heterogeneity in market opinions. However, the results for the DEM/USD and EUR/USD are consistent with this strand of the literature. Overall, our results clearly show that central bank interventions tend to raise the heterogeneity of exchange rate expectations. The outline of the remainder of this paper is as follows. Section 2 presents a discussion of our data set and in Section 3 we present our empirical results. Section 4 contains some conclusions.

In this paper we have investigated the effect of central bank intervention on the heterogeneity of foreign exchange rate expectations using a newly constructed data set. The idea was to investigate one possible implication of the coordination (or microstructural) channel of central bank interventions, namely that official intervention, or, more specifically, secret interventions, can result in an opening up of the range of market expectations. Our measure of individual expectation heterogeneity is based on monthly survey data for two periods: 1992–1994, and 1996–2001. We study the impact of official interventions as reported by central banks (United States, Germany, Eurozone) or by the Ministry of Finance (Japan) and the impact of interventions reported by wire services. We further analyse the differentiated effect of interventions according to whether they have been preceded by rumours or not, and we also study the impact of rumours per se. Although data limitations prevent us from drawing very general results, in terms of time and country samples, our empirical investigation shows that central bank interventions can have a significant impact on heterogeneity at a monthly horizon. Indeed, the heterogeneity of monthly expectations at the 1-, 3- and 12-month horizons is shown to significantly increase in the case of official (especially unexpected) interventions (DEM–EUR/USD) or of expected interventions and “false” rumours (JPY/USD). Hence, central bank intervention can be viewed as able to open up market opinions, albeit in a way which is different for the two markets. The results for the DEM–EUR/USD are consistent with the microstructure literature which suggests that secret interventions may raise informational heterogeneity. The results for the JPY/USD are more mixed in that only expected interventions, which by construction are not secret, seem to matter in addition to rumours.