Communication has become critically important for modern monetary policymakers. The reasons are two-fold: to ensure appropriate accountability for increasingly independent central banks and to improve the management of expectations. However, research has tended to focus more on the effects of the actions of central banks, rather than their communication. In South Africa all previous studies of the extent to which the South African Reserve Bank (SARB) ‘creates news’ or ‘reduces noise’, investigate the movement of market interest rates in anticipation of, or in response to, monetary policy action. In constructing an index to capture the nature of the forward-looking communication in the SARB’s monetary policy statements, this research (Reid and Du Plessis, 2010) extends the South African literature by concentrating specifically on the SARB’s communication.

We focused on the consistency of the communication of the SARB in its monetary policy statements since the adoption of inflation targeting. Consistency is here understood to mean that the SARB’s actions matched its words, in the sense that communication about the likely future path of monetary policy was followed by compatible monetary policy actions. Consistent communication is essential for building the credibility that would improve the SARB’s monetary policy effectiveness and encourage support for monetary policy institutions as a whole.

Creation of a policy inclination index

To facilitate a formal analysis of the informational content of the SARB’s communications, a subjective numerical index (the policy inclination index) was constructed (influenced by Musard-Gies (2005), De Haan and Jansen (2005), Rosa and Verga (2007), Ehrmann and Fratszcher (2007)). Two informed analysts independently read the monetary policy statement released at the end of each monetary policy committee meeting, and allocated an index value between -2 and 2 to describe the inclination toward monetary easing or tightening of the forward-looking components of each statement. This index, capturing the communications of the SARB, was then compared with the SARB’s actual monetary policy actions to judge the consistency of the communication.

The consistency of the SARB’s communication

An exploratory data analysis provided a first indication that the SARB’s communication was consistent. Visual inspection of the evolution of the policy rate and that of the index revealed that the two series have similar patterns, but that the actual policy rate clearly lags the index.

Transition matrices, comparing the index in time t with the policy action in time t+1 were used to classify each policy decision as strongly consistent, mildly inconsistent or strongly inconsistent. The majority of these 60 monetary policy decisions in the sample period were classified as strongly consistent, whereas only 2 of the 60 policy decisions were classified as strongly inconsistent. One of the instances of strong inconsistency was further excluded because STATS SA acknowledged on 12 June 2003 that inflation figures had to be adjusted as they had previously been miscalculated.

The distinction between inflation surprises and policy confusion

An important distinction was made between inflation surprises, where the SARB’s deeds did not match its prior words because the macroeconomic environment changed and policy makers justifiably responded by changing their policy; and policy confusion, where the SARB did not communicate clearly. From a total of sixty MPC meetings, twenty policy decisions were identified as inconsistent or unexpected. These were individually evaluated by referring to the Reuters Econometer forecasts and changes in the Bankers’ Acceptance rates which reflect the extent to which analysts and the financial markets had correctly predicted the policy decisions. Using these criteria, 13 of the 20 decisions were judged not to have been considerable surprises in practice. The seven remaining instances of potential policy confusion were examined in their contexts by referring to media reports and intermeeting communications from the SARB. Three of these were classified as inflation surprises, leaving four occasions that were classified as instances of policy confusion. These results suggested that the SARB had communicated with considerable consistency.

The exploratory data analysis was followed by formal regression analysis. Both OLS and ordered probit regressions again showed that the index does indeed convey information about future changes in the repo rate.

Concluding Remarks

The index was scrutinised in a number of ways, and all these confirmed that the communication of the SARB about the likely future path of monetary policy contained informational content. The communication of the SARB, which is complex in practice, was reduced in this study to a single-figure index. Even this informed simplification of the formal communication confirmed that the SARB’s MPC had succeeded in signalling its likely future policy actions with consistency since the advent of inflation targeting.