Essays in central bank communication and bank lending conditions for firms

Abstract

This dissertation is composed of three independent essays in the broad fields of central bank communication and the transmission of changes in bank lending conditions to firms. The first two essays examine the transmission of monetary policy through words, as central banks communicate with the public, while the third focuses on the more conventional channel of how changes in bank side credit conditions pass through to firms.

The first essay (published in the European Journal of Political Economy) examines the financial market impact of ECB monetary policy press conferences. Using natural language processing methods, in this paper I create two measures of the monetary policy tilt of the ECB for each press conference, indicating how hawkish or dovish the communication is. These two measures closely track interest rate expectations over the tightening and loosening cycle, and provide a particularly useful measure of monetary policy tilt at the zero lower bound. Using intraday data I investigate the impact of changes in the hawkishness or dovishness of the press conference that is free of the impact of the monetary policy decision itself. I show a non-negligible positive (negative) effect on stock prices of a more hawkish (dovish) tone. This indicates that the ECB releases some degree of private information during this press conference, with market participants internalising this communication as being indicative of good (bad) news regarding the future state of the economy rather than indication of future potential increases (decreases) in interest rates. The effect is stronger prior to the introduction of formal forward guidance in July 2013, suggesting that since this period ECB communication has been less surprising to markets, perhaps indicating evidence of more certainty in the future path of rates and perspectives on monetary policy between that date and the end point of the study in 2020.

The second essay builds on themes in the first essay by also examining topics from central bank communication, but this time turning to the temporal aspects of communication rather than the tone. It presents an additional explanation for why central bank communication moves markets beyond the simple release of private central bank communication. This paper posits that markets move due to an ?information deficit? that is held by the public that the central bank helps to ?fill in? using its communication. We posit that there are three channels through which this information deficit can be filled in, namely through the central bank providing (i) an updated evaluation regarding the state of the economy (ii) a different projection from the current state to the likely future evolution of the economy and (iii) differences in how the policymaker responds to the assessed economic outlook. Using NLP methods we produce measures of the temporality of central bank communication and illustrate that both communication about the past and future significantly improve the ability to explain the news in yields around both the Federal Reserve and ECB policy statements compared to pure topic based measures. We surmise this is evidence of both the projection and evaluation channels at play. Moreover, to further illustrate the importance of the information deficit, we show that speeches generate greater news effects through moving yields when the information deficit is greater following the press conference, as measured using the similarity of speeches and responses to questions. From a policy perspective we suggest that these results make the communication of a single fixed reaction function very difficult even if it is highly desirable due to changes in the assessment of what is driving underlying changes.

The third essay examines a more conventional approach to assessing the impact of retail bank behaviour on the broader real economy by examining how changes in credit standards by individual banks feed through to firm outcomes. Using a novel dataset linking firm level data from the Survey on Access to Finance for Enterprises (SAFE) and bank level data from the Bank Lending Survey (BLS) we find that tighter credit standards decrease loan availability as reported by firms, increase the likelihood they report access to finance as the worst problem they are facing and decreases their investment activity. This effect fades when adding country-sector-time fixed effects, suggesting that this aggregate effect is more indicative of cyclical conditions, however the results become significant once interacting with firm-level demand as proxyed by their reported change in need for bank loans. We find that results are asymmetric, with stronger effects for a tightening than an easing, while a more diversified array of funding sources insulates firms from changes in credit standards. This illustrates the importance of investigating both demand and supply side behaviour when investigating changes in credit standards and has implications for understanding the bank lending channel of monetary policy. We use these micro results to propose a new measure of demand adjusted credit standards that can be used to analyse broader credit dynamics.