Effects of transparency, monetary policy signalling and clarity of central bank communication on disagreement about inflation expectations

ABSTRACT

The literature on transparency and central bank communication and the literature on disagreement about expectations are evolving; however, both have been evolving separately. Despite the advances in the literature, several key issues remain open and there are gaps to be filled. Therefore, this study analyses the effects of monetary policy signalling and clarity of central bank communication on disagreement about inflation expectations. It also investigates whether greater transparency coincides with lower levels of disagreement in inflation expectations in Brazil. The findings suggest that transparency is important to reduce disagreement about inflation expectations. Moreover, our estimates indicate that central bank communication and clarity affect disagreement about inflation expectations in Brazil.

KEYWORDS:

• Transparency
• communication
• clarity
• disagreement
• central bank

JEL CLASSIFICATION:

• E44
• E58
• E37

Notes

¹ We consider the expectations formed by financial market experts. According to Ullrich (2008), financial market experts watch the central bank very closely; we would expect that an influence on their expectation formation is more likely than an influence on expectations of consumers.

² The Michigan Survey of Consumer Attitudes and Behaviour, the Livingston Survey and the Survey of Professional Forecasters.

³ Australia, China, Hong Kong, India, Indonesia, Japan, Korea, Malaysia, New Zealand, Philippines, Singapore, Thailand. The primary source is compiled by the consulting firm Consensus Economics.

⁴ Canada, France, Germany, Italy, Japan, United Kingdom and United States.

⁵ http://www.bcb.gov.br/?TIMESERIESEN.

⁶ https://www3.bcb.gov.br/expectativas/publico/consulta/serieestatisticas.

⁷ http://www.bcb.gov.br/?MINUTES.

⁸ Focus-Market Readout presents the results of Gerin’s market expectations survey, a daily survey of forecasts of roughly 120 banks, asset managers and other institutions (real sector companies, brokers, consultancies and others) for the Brazilian economy. The survey of market expectations began in May 1999 as part of the transition to the inflation targeting system. The survey was created in order to monitor market expectations regarding the main macroeconomic variables, and improve the inputs for the monetary policy decision-making process. The system calculates sample statistics from the data gathered in real time, and thereby enables the generation of daily reports for the members of the Board. One of these reports, the Focus-Market Readout, is published every Monday at the Central Bank website at the internet.

⁹ This instant is characterized by a specific date, namely, a day d, a month m and a year a.

¹⁰ The number of agents in $\mathrm{I}$ is I.

¹¹ j = 0: current year; j = 1: next year immediately after the current year; j = 2: two years after the current year; j = 3: three years after the current year; j = 4: four years after the current year.

¹² It is worth mentioning that the Gerin informs to the public only the main statistics of the distributions of expectations, that is, means, medians, minimum and maximum values, SDs and coefficients of variation.

¹³ An example could help to clarify this issue. Suppose that an agent, in March 2000, computes his expectation about the value of the inflation rate in the end of 2000. In this case we can say that the time horizon of the forecast is 10 months because the first 2 months of 2000 have already passed and inflation figures for January and February are known. By the same line of reasoning, when this agent computes his inflation expectation in September 2000 about the value of the inflation rate at the closing of 2000, the time horizon of his forecast decreases to only 4 months.

¹⁴ Indeed, the disagreement measure observed in March 2000 for the value that the inflation rate will take in the end of 2000 tends to be greater than the disagreement measure observed in September 2000 for the value that the same variable will take at the closing of 2000. The divergence measure tends to increase again in March 2001, since the current year becomes 2001 and the time horizon of the forecast becomes 9 months.

¹⁵ For example, the weighted inflation target in June of 2003 is a result of (6 × target 2003 + 6 × target 2002).

¹⁶ For example, the weighted inflation target in July 2004 is a result of (5 × target 2004 + 7 × target 2005).