An empirical analysis of monetary policy communication in Nigeria

The thesis examines the monetary policy communication of the Central Bank of Nigeria (CBN)between 2001 and 2018 by: (a) probing the clarity of the Monetary Policy Communiqués; (b)measuring communication and ascertaining the transmission effects; (c) scrutinizing the intensityof fiscal policy communication; and (d) gauging the public’s knowledge of monetary policy.The Flesch Kincaid and Gunning Fog readability tools were used to measure clarity, while themultiple regression, Impulse Response Function and Variance Decomposition techniques wereapplied to test how clarity responded to changes in macro variables in the period. The mainfindings are: (a) The number of paragraphs, sentences, and words of the ccommuniqués increasedin the period, suggesting improved transparency by the CBN; (b) The high Flesch Kincaid andGunning Fog scores signifies that the communiqués were complex, buttressing the findings ofTumala and Omotosho (2019) that CBN’s statements have become more difficult.To measure CBN’s communication and the transmission effects, two indicators were developed -the Monetary Policy Stance (MPS) that codes the tightening (+1), hold (0) and loosening (-1)postures as widely applied in the literature; and the Monetary Policy Communication on Inflation(MPCI) that reflects statements on inflation in line with the ECB KOF approach. The empiricalscrutiny followed Neuenkirch (2013) that included the KOF Communicator in a transmissionmechanism study for the ECB. Also, CBN’s study on monetary policy transmission in 2010provided guide in terms of the estimated equations, which were modified with the MPS and MPCI.The findings from the graphic analysis of the MPS showed that the CBN tightened its stance in 26or 37% of the communiqués while it was 11 or 16% for easing and 33 or 47% for the neutralposture. The average MPCI of 0.135 indicates that the CBN communicated more about upside riskto inflation, while the relationship between voting to raise policy rate and the MPCI is positive butweak, suggesting that the CBN also used other tools to manage inflation in the period.A juxtaposition of the movement between MPCI and policy rate showed a paradox as the upsiderisk sentiments to inflation coincided with falling or flattening policy rate in most parts of theperiod reviewed. The likely reasons are: (a) Given the multiple objectives and policy instruments,the CBN sometimes alter other tools instead of raising the policy rate; (b) CBN probably sawinflationary threats as temporary in those periods and decided to be neutral on policy; (c) Thepolicy rate is the most visible tool to politicians who see raising rates as undermining theirpreference for growth, making the CBN to sometimes adjust other tools instead of the policy rate.By means of the ARDL/Bounds testing for cointegration and the Error Correction Model, the keyresults from the empirical scrutiny of the effect of communication on the transmission process are:(a) The Error Correction Terms (ECTs) are negative and significant in all the models, implyinglong-run causality; (b) The ECTs are below 20% in all the models, indicating slow speed ofadjustment in models with or without communication. This suggests that communication did notimprove transmission of monetary policy between 2001 and 2018. Given that communication isaimed at managing expectations, the inference is that the result underpins that of CBN’s 2010analysis which found weak existence of policy transmission through the expectations channel; (c)The models with MPCI that focus on inflation performed relatively better than the ones with MPSthat reflects the multiple objectives and tools; (e) The short-run causality results reinforce theimportance that the CBN attaches to monetary targeting as money supply granger-causes GDP andinflation. Also, the CBN’s tightening, neutral and easing postures as captured by the MPS isimportant for short-run macroeconomic management given that the MPS granger-causes GDP.To analyze the intensity of CBN’s communication about fiscal policy, a template was designedand used to develop a Fiscal Policy Communication Indicator (FPCI) after the method of Allardet al. (2013). The multiple regression was used to estimate how macro-fiscal variables affectedcommunication about fiscal policy, while controlling for other effects, also in line with Allard etal. (2013). The estimations were however complemented with the Impulse Response and VarianceDecomposition methods that are widely utilized in the central bank communication literature.The results are as follows: (a) A total of 409 statements and 20,193 words or 11.8% of total wordcount of all the communiqués were used to discuss fiscal policy; (b) CBN communicated morepositively about fiscal policy when the oil-price based fiscal rule was introduced. The formationof the Sovereign Wealth Fund resulted in more normative statements as the posture changed fromcommendation to advisory; (c) Fiscal deficit and inflation have significant effects on the FPCIwhile effect of public debt and money supply are insignificant; (d) The variance decompositionresults showed that 77% of the Forecast Error Variance in the FPCI is due to own effect, meaningthat the explanatory variables did not significantly influence the variance in the FPCI; (e) FPCI isgranger-caused by fiscal deficit, public debt and money supply, while it is not granger-caused byinflation. The combined effect of all the indicators showed that they all granger caused FPCI.To examine the public’s knowledge of monetary policy, the questionnaire by Cruijsen et al. (2015)was modified and used for data gathering while the logistic regression method was used for theanalysis. The results are: (a) Knowledge of monetary policy is not affected by gender; (b) Age isvital as older respondents are more conversant about monetary policy; (c) Educationalqualification does not increase the chance of being well-informed about monetary policy; (d) Realsector respondents tend to be more familiar compared with those in other sectors; (e) Respondentswith desire to be well-informed are more knowledgeable about monetary policy; (f) Workexperience increases the likelihood of being familiar about monetary policy; (g) Higher incomeearners are well-informed about monetary policy; (h) Users of Newspapers, CBN’s website, Radioand Television tend to be more aware about monetary policy; (i) Using current information, pastinformation or both, does not impact the likelihood of being conversant about monetary policy.Overall, the study concludes that:a) CBN’s communication of monetary policy through the policy communiqué is complex.b) The transmission mechanism of monetary policy was not enhanced by communication inthe review period.c) The discussions about fiscal policy have expectedly increased over the years, with theCBN’s positive communication increasing in periods of strategic fiscal policy initiatives.d) The public’s knowledge of monetary policy is enhanced by age, income and the mediareporting through Newspapers, Television and Radio.The study recommends as follows:a) To improve clarity, CBN should devise means of explicitly conveying its multipleobjectives and tools and also simplify its vocabularies and choice of words.b) The transmission process can also be enhanced if clarity improves with simplification ofvocabularies and choice of words. This will reduce the mixed signals to the public.c) Commending government when certain actions are taken and advising them to take firmmeasures in some instances will improve monetary-fiscal bond.d) The press briefing channel of communication should be strengthened because it is effectivein helping the media to convey the policy messages clearly to the public.