اعتقادات مقامات بانک مرکزی در مورد مداخله ارزی

This paper presents the results of a survey of monetary authorities with respect to foreign exchange intervention. The survey offers evidence on new issues that would otherwise be difficult to investigate, such as response times, non-foreign exchange factors in intervention and profitability. The survey also reveals new evidence on previously studied issues, such as channels of effectiveness. Respondents disagreed with predominant views on intervention and volatility and common arguments against intervention. Exchange rate regimes explain central bank beliefs about important aspects of intervention, including factors that lead to detection of secret interventions and the potential profitability of intervention.

The frequency of foreign exchange intervention among developed nations has waned in the last 10 years. The Japanese authorities, for example, have not intervened since March 2004. Even as the frequency of developed country intervention has declined, intervention sizes have increased and intervention remains a policy tool for both developed and developing countries. The study of the causes and effects of intervention remains pertinent, therefore, both for its own sake and as a way to shed light on microstructure issues such as the role of information transmission. The paucity of accurately timed intraday data on intervention and the simultaneity of intervention and exchange rate returns complicate the econometric study of intervention, however (Neely, 2005a, Neely, 2005b and Fischer, 2006). The views of individuals familiar with the practice of intervention provide a natural complement to econometric inference or event studies. Central bankers who conduct intervention have collectively witnessed thousands of natural intervention experiments and presumably have important insights into its workings. This observation has motivated at least three surveys of central bankers on the subject of intervention: Neely, 2000 and Mihaljek, 2004 and Lecourt and Raymond (2006). These three surveys differed in their coverage and emphasis. Neely (2000) surveys the foreign exchange desk/reserve management departments of 22 central banks, asking about the mechanics of intervention. Most responding authorities conduct intervention in spot markets, with domestic commercial banks, during domestic business hours. Misalignments and volatility motivate interventions, while desire for market impact produces mixed effects on secrecy. Interestingly, Neely (2000) finds that central banks unanimously support the idea that intervention is effective in changing exchange rates. Lecourt and Raymond (2006) follow up by surveying only central banks of industrialized countries, exploring beliefs about the effectiveness of intervention through various channels. After emphasizing the importance of expectations and credibility to the effectiveness of intervention, the authors go on to describe the quantity and frequency of G3 intervention from publicly released data. Mihaljek (2004) exclusively focuses on authorities of developing countries, finding that interventions are small relative to market size and that most authorities view intervention as effective in calming disorderly markets. Mihaljek (2004) finds that respondents consider intervention to work through expectations of both future monetary policy and intervention. The study concludes that intervention's effectiveness depends on the consistency of macro/monetary fundamentals with intervention. Finally, a substantial part of the work deals with the beneficial effects of reserve accumulation on sovereign credit ratings and vulnerabilities to external shocks. Instead of focusing on intervention mechanics, the present survey asks market participants about their beliefs about the motivations for, effects of, and arguments against intervention. Neither does survey participation require recent intervention, the cover letter specifically asked authorities to refer to past experiences if they have conducted little or no recent intervention. Several authorities declined to answer the survey, however, citing a lack of recent intervention experience and/or institutional memory. In addition to studying authorities' beliefs on a wide variety of intervention topics, the present survey considers whether a country's exchange rate regime and per capita output explain the responses. As discussed earlier, the exchange rate regime can obviously influence the motivations for and outcome of intervention. Per capita output serves as a rough proxy for the sophistication of financial markets. Central banks that face deep and sophisticated financial markets might well have different views on the reasons for and efficacy of intervention than those facing developing markets. To presage the results, central bankers' experiences with exchange rate regimes and – to a much lesser extent – the level of development of their financial markets are often correlated in sensible ways with their beliefs about intervention. A number of papers have surveyed foreign exchange traders, including at least three that touch on beliefs about intervention. Cheung and Wong, 2000 and Cheung and Chinn, 2001, and Cheung et al. (2004) survey traders in Asia, the United States, and the United Kingdom, respectively. None of these surveys encounter more than modest support for intervention among traders. Most traders believe that intervention raises volatility and are split about whether intervention achieves its goals or moves exchange rates toward fundamentals. The most positive response was in Cheung and Wong (2000). From 55% (in Tokyo) to 71% (in Hong Kong) of traders say intervention moves exchange rates toward fundamental values but only 32–61% believe that intervention achieves its goals. It is worth noting that papers that ask central bankers about their intervention practices and beliefs – like the present paper – face different challenges and contribute a very different perspective than papers that query foreign exchange traders. The latter papers spend a good deal of time inquiring about job titles, seniority and position limits of their respondents in the belief that these characteristics explain beliefs about the foreign exchange market. In contrast, the present study surveys people with similar job titles and a fair amount of seniority, which makes it unlikely that these factors determine differences in their views on intervention. The current survey contributes to the intervention literature by broadly surveying the beliefs of central bankers who have been directly involved in foreign exchange intervention and/or reserve management. It expands upon some previously considered issues, such as secrecy and factors in successful interventions. For example, this survey asks about factors that lead to the detection of secret interventions. But it also extends the methodology to new topics, including effects on volatility, intervention response time, non-foreign exchange motivations for intervention, the potential for profitable intervention and arguments against intervention. Issues such as response time or non-foreign exchange motivations for intervention would be almost impossible to investigate with other methods but are of considerable importance for econometric investigations. For example, high-frequency studies to investigate the effects of intervention are immune to endogeneity bias only if the frequency of the econometrician's data is greater than the response time of the intervening authority. Similarly, non-foreign exchange factors that influence intervention are potential instruments in a study of intervention. It is also useful to emphasize the limitations of any such survey. While it might be desirable to ask a wide range of questions on intervention practices and interaction with monetary and other policies, a long survey would doubtless reduce the response rate and sample size. To maximize the response rate and sample size, the survey was kept concise (two pages) and omitted many questions designed to catalogue specific intervention practices that Neely, 2000 and Mihaljek, 2004, or Lecourt and Raymond (2006) covered. Keeping the survey short helped produce a very high response rate (45%) compared to overall response rates in the range of 5–20% for Cheung and Wong, 2000 and Cheung and Chinn, 2001 and Cheung et al. (2004). Another potential limitation is that central bankers might have incentives to answer some questions dishonestly. For example, central bankers might be reluctant to report that intervention raises volatility. Systematic dishonesty seems unlikely for several reasons. First, the responses are anonymous, meaning that no central banker must justify or explain a potentially embarrassing response. Second, practically speaking, no central banker will be personally affected by the marginal effect of his/her own anonymous answers published in an academic study. And third, many central bankers – even those who have been involved in intervention – will publicly criticize past intervention practices or oppose the practice itself.1 Therefore, it seems likely that the respondents truthfully disclose their views.

This paper describes the results of a survey sent to 52 monetary authorities regarding their experiences with and beliefs about foreign exchange intervention. Twenty-three of those authorities responded to some or all of the questions, including 17 with experiences exclusively in floating rate regimes and six that have experience in fixed exchange markets. Most responses came from heads of departments devoted to monetary and/or foreign exchange operations. The purpose of the survey was twofold: (1) to investigate questions that would be very difficult or impossible to investigate with more conventional empirical techniques; and (2) to challenge the results of the empirical literature on intervention with new evidence from participants. The paper provided evidence on several issues that have been little researched – perhaps because they would be very difficult or impossible to investigate with traditional methods. For example, this paper established that reaction times for intervention are fairly rapid, but time-varying, dependent on whether the authority is alert to market developments. The paper also inquired about non-foreign exchange factors in intervention decisions. All the factors cited – wars, assassinations, elections etc. – are unlikely to directly elicit intervention, but rather do so through their effects on the foreign exchange market. Further, the paper studied the determinants of the size of intervention, which Kearns and Rigobon (2005) emphasize is a separate empirical issue from the propensity to intervene. The paper finds that monetary authorities are unlikely to reinforce successful intervention. And the paper extended the literature on secrecy by inquiring about the factors in the probability of detection of secret interventions. Reassuringly, the central bankers supported many of the conclusions of empirical research on intervention. For example, respondents agreed that intervention's effects are not confined to the market in which it is conducted and the signaling, coordination and liquidity channels are more plausible than the portfolio balance channel. And they found Beine and Bernal's (in press) motivations for secrecy to be plausible. On other issues, however, correspondents dissented from academic conventional wisdom. Respondents were likely to disagree with the assertion that intervention increases volatility, for example. And central bankers who conduct intervention were unpersuaded by most of the common arguments against intervention. The only argument that participants tended to support is that intervention might be used to substitute for other necessary policy changes. Despite the small sample size, exchange rate regime and national income sometimes significantly explain the attitudes and beliefs of authorities on intervention. For example, per capita GDP significantly explains whether a responding authority believes that substantial resources are important for the success of intervention. Floating rate authorities consider the size of intervention to be more likely to raise the probability of a successful intervention. And fixers are more likely to believe that coordination increases the probability that the market will detect secret interventions. Monetary authorities are a bit schizophrenic about the profitability of intervention. The mean response to the statement that “active reserve management could generate a positive excess return – in excess of opportunity cost – at a reasonable risk for the taxpayer”, was only 2.95 (Table 5, row 1), indicating very mild disagreement, on average. But central bankers agreed that their trading desks could trade profitably at intraday horizons through multi-year holding periods. Floating authorities were more optimistic than fixers about their ability to generate excess returns at reasonable risk and the difference is marginally statistically significant, despite the very small samples. The discrepancy between the ability to actively manage reserves and the ability to trade profitably might be due to the reluctance of central bankers to take on such a potentially distracting responsibility, despite the fact that they believe that they could do so successfully.