Across the world, considerable attention is being paid to analysing central bank communication, especially in the aftermath of the 2008 Global Crisis. Monetary policy communication is regarded as an effective tool in the process of anchoring agents’ inflation expectations and improving monetary policy transmission (Keida and Takeda 2020, Woodford 2005). This is especially true of countries that have adopted inflation targeting as a monetary policy framework because the success of such a regime hinges on the effectiveness of transmission of monetary policy announcements.

In this column, we analyse how the monetary policy communication of the Reserve Bank of India (RBI) has changed from 1998 to 2018, across the regimes of five different governors, especially after the adoption of inflation targeting and what may have been the effect of different aspects of the RBI’s communication on financial markets. As a new inflation targeting country where conventional channels of monetary transmission do not work well, India is an excellent case study for quantitatively analysing the central bank’s monetary policy communication and exploring the transmission thereof to financial markets.

Specifically, we ask three questions (Mathur and Sengupta 2020). First, does the de-jure move to inflation targeting get reflected in the manner in which the RBI communicates its monetary policy decisions? We look for the most frequently used words in the monetary policy statements of pre-inflation targeting and post-inflation targeting regimes and visualise them in word clouds.

Second, how has the linguistic complexity of monetary policy statements of the RBI evolved over the last two decades? To explore this, we use two indicators: length and readability of monetary policy statements.

Finally, what is the effect of the RBI’s monetary policy statements on the financial markets? We hypothesise that lengthier or more complex statements induce greater volatility of returns in the financial markets. Apart from being cognitively taxing on the reader, unclear communication increases the likelihood for market participants to diverge in their interpretation of the information conveyed, thereby resulting in greater volatility (Ehrmann and Fratzscher 2007, Jansen 2011, Atmaz and Basak 2018, Weidmann 2018).

Our study is closely related to Smales and Apergis (2017) who analyse the complexity and readability of Federal Open Market Committee (FOMC) statements and find that lengthier and more complex statements result in greater volatility and trading volumes in the US financial markets.

Before and after inflation targeting: Length, readability and content

Our study covers the period from October 1998 to June 2018. During this period, the RBI followed a multiple-indicator approach from 1998 to 2016, and an inflation targeting regime from 2016 to 2018. Governors Bimal Jalan, Y.V. Reddy, D. Subbarao, and Raguram Rajan belonged to the first era, while Urjit Patel belongs to the second era.1

We measure length as the number of words used in each statement. The main idea is that longer statements might be more deterring for readers requiring higher costs of information-processing (Li 2008).

Figure 1 Evolution of the length of RBI’s monetary policy statements

Note: This graph shows the two-statement rolling average of the length of statements as measured by the number of words across the regimes of the five governors. MPC indicates the official advent of inflation targeting.

Figure 1 shows that prior to the adoption of inflation targeting in 2016, the RBI’s monetary policy statements were on average lengthier compared to those issued in the post-inflation targeting era. In the pre-inflation targeting regime, the average monetary policy statement was roughly 13,000 words; since inflation targeting adoption, this has fallen by three-quarters to 3,084 words. The statements of the Monetary Policy Committee (MPC) are still roughly six times longer than the monetary policy statements of both the FOMC and the ECB.

READ ALSO  Fear and employment during the COVID pandemic

Lengthier statements might still be easy to read, so we complement our analysis by using a standard index of readability, the Farr-Jenkins-Paterson index (Farr et al. 1951), which counts the number of one syllable words per 100 words.2 The idea is to be able to quantify the complexity and lexical diversity of the statements. Lower values of the index indicate lower readability. We find that the readability of RBI’s statements is fairly low on average but has improved with the advent of the inflation targeting regime (Figure 2).

Figure 2 Readability of RBI’s monetary policy statements

Note: This graph shows the evolution of the average FJP readability index (vertical axis) across the regimes of the five governors. The index is negative with lower index values implying lower readability. The horizontal axis measures the number of statements per governor.

There is some evidence that the shortening of the statements and the improvement in readability may have partly been a function of governor-specific factors as well, and not just of the shift to inflation targeting alone. For instance, there was a marked decline in the length of statements when Subbarao became the governor and again from the time when Rajan took office in 2013, as can be seen in Figure 1. Likewise, as shown in Figure 2, readability of the monetary policy statements issued during Subbarao’s time was on average higher than the previous two governors’ regimes and in fact similar to that of the MPC’s.

Next, we use word clouds to uncover the implicit focal variables for the inflation targeting and pre-inflation targeting regimes. We follow the standard steps for processing textual data and generate word clouds where the size and colour of the words is directly proportional to their frequencies (Figure 3). The hypothesis is that the MPC’s word cloud would be dominated by the words ‘inflation/prices’ and related words, while those of the previous regimes might not be.

Figure 3 Word clouds, before and after inflation targeting

a) pre-MPC (Oct 1998-Oct 2016)

b) post-MPC (Oct 2016 – Jun 2018)

Note: This graph shows word clouds for the pre-MPC (Oct 1998–Oct 2016) and post-MPC regimes (Oct 2016–Jun 2018), where Oct 2016 refers to the first official meeting of the MPC. The colours and sizes of the words are proportional to their raw frequencies. The 40 most frequent words are plotted in each picture.

We see that in the pre-MPC word cloud, the words inflat* or price are not the most prominent ones. Instead words such as financi*, market, credit etc., appear to be used more frequently implying that during this period the RBI governors were focussing on multiple factors such as credit growth, financial conditions, etc when deciding the policy rate. Also worth noting is the occurrence of the word exchang* in the word cloud of this period. This hints at the RBI’s concern about exchange rate volatility. It does not, however, appear in the MPC word cloud.

In contrast, in the MPC word cloud, the most prominent words are inflat*, price, and growth. This implies that the focus of the statements has evidently shifted towards more inflation-related terms (such as food, cpi, demand, crude, and oil) in keeping with the inflation targeting mandate.

READ ALSO  The US excess mortality rate from COVID-19 is substantially worse than Europe’s

Impact on financial markets

We analyse the impact of length and readability of the monetary policy statements on the volatility of returns in the equity, currency, and bond markets. As a proxy for equity markets, we use daily data on the Nifty50 index, which captures the 50 most liquid stocks in India. For currency markets, we use daily returns on the Indian Rupee-US Dollar pair, while for bond markets we restrict attention to the ten-year government bond yields, which is the most liquid tenor. We calculate the returns over a seven-day period starting from the day of the monetary policy announcement (and hence publication of the statement on the RBI’s website).

Our hypothesis is that an increase in the length of a statement or a decrease in the readability should increase the volatility of returns. The longer the statement and lower the clarity of information conveyed, i.e. the lower the signal-to-noise ratio, the greater the scope for market participants to diverge in their beliefs or opinions about the current and future path of policy. This in turn induces greater volatility in returns.3

To account for the possibility that a statement containing a monetary policy surprise may also drive up financial market volatility, we control for any unanticipated changes in the policy rate by using the absolute difference in the Overnight Index Swap (OIS) rate, between t-1 and t, with t being the day of the monetary policy announcement.

In our estimation strategy we address the potential concern that an already complicated or worsening macroeconomic situation drives both market volatility, as well as length of the monetary policy statement. In addition, we account for the possibility that the response of market activity to monetary policy related news may depend on the stage of the business cycle, i.e. whether the economy is in a recession.

We find that after controlling for a host of factors, a 1% increase in the number of words (a rough increase of about 115 words) is strongly correlated with a higher stock market volatility of roughly 0.24% and a rise in currency market volatility of 0.23%. We do not find any effect on bond market returns. These results are in line with the existing literature (Jansen 2011, Smales and Apergis 2017). We also do not find any significant effect of readability on returns volatility. One reason could be that there is not much variation in the readability of the statements over the sample period, especially as compared to other central banks such as the FOMC.

Conclusion

A growing strand of literature has been examining the impact of the quality of central bank communication on financial market volatility. In this column we have presented one such study that complements the findings of this literature and unlike the existing papers, does so in the context of an emerging country central bank that has only recently transitioned to an inflation targeting regime.

Quantifying the monetary policy communication of the Reserve Bank of India and analysing its evolution over a 20-year period, we find that the move towards an inflation targeting regime is reflected in the monetary policy statements of the RBI. With the advent of inflation targeting, RBI’s monetary policy communication seems to have improved significantly. We shed light on the possibility that transmission may get impeded if the RBI’s monetary policy communication is not concise or clear.

READ ALSO  How GDPR affects global markets for data

This kind of a study will help throw light on the optimal communication strategy that an emerging economy central bank such as the RBI can devise in order to effectively influence agents’ inflation expectations and achieve its official objective of controlling inflation.

References

Atmaz, A and S Basak (2018), “Belief dispersion in the stock market”, The Journal of Finance, 73(3): 1225–1279.

Blinder, A S, M Ehrmann, M Fratzscher, J de Haan and D J Tansen (2008), “Central bank communication and monetary policy: A survey of theory and evidence”, Journal of Economic Literature 46(4): 910–945.

Ehrmann, M and M Fratzscher (2007), “Communication by central bank committee members: Different strategies, same effectiveness?”, Journal of Money, Credit and Banking 39(2-3): 509–541.

Farr, J N, J J Jenkins and D G Paterson (1951), “Simplification of Flesch reading ease formula”, Journal of Applied Psychology 35(5): 333–337.

Geraats, P (2002), “Central Bank Transparency”, Economic Journal 112(483): 532–565.

Jansen, D J (2011), “Does the clarity of central bank communication affect volatility in financial markets? Evidence from Humphrey-Hawkins testimonies”, Contemporary Economic Policy 29(4): 494–509.

Keida, M and Y Takeda (2020), “The Art of Central Bank Communication: Old and New”, VoxEU.org, 17 April.

Li, F (2008), “Annual report readability, current earnings, and earnings persistence”, Journal of Accounting and Economics 45(2-3): 221–247.

Mathur, A and R Sengupta (2019), “Analysing monetary policy statements of the Reserve Bank of India”, IHEID Working Papers 08-2019, May 2019.

Mishra, P, J P Montiel and R Sengupta (2016), “Monetary transmission in developing countries: Evidence from India”, IMF Working Papers No. WP/16/167, International Monetary Fund.

Smales, L and N Apergis (2017), “Does more complex language in FOMC decisions impact financial markets?”, Journal of International Financial Markets, Institutions and Money 51: 171–189.

Weidmann, J (2018), “Central bank communication as an instrument of monetary policy”, Lecture by Dr. Jens Weidmann, President of the Deutsche Bundesbank and Chairman of the Board of Directors of the Bank for International Settlements at the Centre for European Economic Research, Mannheim, 2 May 2018.

Woodford, M (2005), “Central bank communication and policy effectiveness”, in The Greenspan Era: Lessons for the Future, Federal Reserve Bank of Kansas City, pp. 399–474.

Endnotes

1 While inflation targeting was formally operationalised in October 2016 during the tenure of Governor Urjit Patel and a monetary policy committee was appointed to decide the policy rate going forward, RBI has been implicitly following an inflation targeting framework from February 2015 onwards under the Governorship of Raghuram Rajan.

2 The most commonly used readability indicator in the literature is the Flesch-Kincaid grade level index which gives the number of years of US education required to read and understand a text. However, this index is not useful for our case. India is not a native English-speaking country. When we apply this index to the RBI’s monetary policy statements, we find that it is unequipped to pick up the variation in communication across the Governors: the range of the index is only 1.8 years, from 14.7 years to 16.5 years, over a 20 year period. This tells us that the statements are complex on average, but not how their complexity has changed over the years, which is our primary focus.

3 See for example: Geraats 2002, Ehrmann and Fratzscher 2007, Blinder et al. 2008, Atmaz and Basak 2018, Weidmann 2018, Jansen 2011.

Via VOX EU