Jeremy Chiu, Richard Harris and Evarist Stoja
Financial market shocks: do they matter for the economy?
Financial markets are intrinsically volatile, constantly fluctuating in response to a wide variety of news. Often, these shocks to volatility are short-lived, perhaps reflecting a one-off adjustment in asset prices or the market’s overreaction to news, and have a tendency to dissipate rapidly. But sometimes they lead to a sustained increase in market volatility, reflecting a deeper uncertainty over the future macroeconomy that can take time to resolve itself. Indeed, a considerable body of empirical evidence suggests that financial market volatility is made up of two components: a slowly varying ‘core’ component and a ‘transitory’ component that dissipates quickly. We develop a way to identify each type and estimate how they affect the broader economy.
Core and transitory volatilities
In our recent paper, we use a cyclical volatility model to decompose U.S. equity market volatility into its core and transitory components. We first extract from daily volatility (which we proxy by absolute daily returns) the core component by applying a one-sided Hodrick-Prescott filter with a standard value of the smoothing parameter to a rolling window of the past 500 observations. We sum the daily core component of volatility over each day in the month to yield the core component of monthly volatility. The transitory component of monthly volatility is then computed as the difference between total monthly volatility and the core component derived in the first step. Figure 1 displays volatility as well as its two components.
Panel 1: Core and Total Volatilities
Panel 2: Transitory Volatility
Notes to Figure 1: Monthly total volatility (black), core volatility (blue) and transitory volatility (green) of the (log annualised) daily returns for the U.S. stock market. The sample period is 02/01/1990 to 30/6/2015.
Core volatilities matter more for the macroeconomy
To examine the link between financial market volatility and the wider U.S. economy, we estimate Structural Vector Autoregression (SVAR) models using the two volatility components, and several measures of the macroeconomy , including the industrial production growth rate, the inflation rate and the short-term interest rate. We also include Shiller’s crash confidence index as a proxy for investor sentiment where a low value of the index corresponds to low investor sentiment and vice-versa. Our data sample is July 2001 to June 2015. We investigate the bidirectional relationships between equity market volatility and the macroeconomic variables in order to establish (a) how macroeconomic shocks impact the different components of volatility, and (b) what, if any, is the impact of the different components of volatility on the macroeconomy. We identify three structural shocks related to aggregate demand, aggregate supply and monetary policy. We also identify structural shocks to volatility and investor sentiment.
Our first result is that adverse Aggregate Demand and Aggregate Supply shocks both create a significant and sustained increase in both core and total volatility, with the former peaking later and staying significant for a longer period. For illustration, Figure 2 presents the impulse responses arising from adverse aggregate demand and supply shocks. The responses of core volatility (the blue line) last for longer compared with those of total volatility (the black line). In contrast, we do not find any significant change in transitory volatility (the green line) conditional on these macroeconomic shocks.
Notes to Figure 2: Impulse Responses for adverse Aggregate Demand and Aggregate Supply shocks computed by the SVAR model using U.S. stock market volatility. In each case, the black line corresponds to the system using total volatility, the blue line corresponds to the system using the core volatility and the green line corresponds to the system using the transitory volatility. Error bands have been omitted to improve readability. The sample period is July 2001 – June 2015. See Chiu, Harris, Stoja and Chin (2016) for details.
Our second result is that shocks to core volatility have a deeper recessionary impact than total volatility shocks. Figure 3 reveals the following: given the same shock to core volatility (the blue line) and total volatility (the black line), the former leads to a more sustained rise in volatility and a deeper contraction in economic activity. In contrast, shocks to transitory volatility (the green line) create short-lived and mostly insignificant macroeconomic responses.
Notes to Figure 3: Impulse Responses for adverse volatility shocks computed by the SVAR model using U.S. stock market volatility. In each case, the black line corresponds to the system using total volatility, the blue line corresponds to the system using the core volatility and the green line corresponds to the system using the transitory volatility. Error bands have been omitted to improve readability. The sample period is July 2001 – June 2015. See Chiu, Harris, Stoja and Chin (2016) for details.
We also investigate how the two volatility factors interact with changes in investor sentiment. We consider changes in investor sentiment (as opposed to the level in our SVAR model above) because the theoretical literature shows that sentiment-based decisions of uninformed investors may lead to excess volatility. We find evidence that changes in investment sentiment are associated with both volatility factors. However, the link with transitory volatility is stronger. Owing to space constraints, we do not report our results here; we refer the interested reader to our paper for a full description of the results.
In a separate exercise, we also explore the dynamics of total volatility and its core and transitory components for the FTSE All Share, FTSE 100 and FTSE 250 indices over the period 1 January to 30 September 2016. As figure 4 shows, three notable features are evident. First, the EU referendum of 23 June led to a large increase in the volatility of the FTSE 250 index, but had a much smaller impact on the FTSE 100 and FTSE All Share indices. Second, almost all of the increase in the volatility of the FTSE 250 was transitory in nature, which – given our finding above – suggests that it reflected mainly investor sentiment rather than macroeconomic conditions. Third, there was a small but sustained rise in the core volatility of the FTSE 250 index ahead of the EU referendum, but this has since started to decline. In contrast, the core volatility of the FTSE 100 and FTSE All Share Indices has been declining since the start of 2016.
Notes to Figure 4: The figure shows the total volatility of the FTSE All Share, FTSE 100 and FTSE 250 indices as well as its core and transitory components, from 1 January to 30 September 2016.
Policy Implications and Conclusions
Our results carry important policy implications. Firstly, our volatility decomposition provides policy makers and practitioners with a more precise measure of the underlying long-term volatility in financial markets, free of the transitory component that reflects short-term adjustments and investor sentiment. Secondly, we have shown that this core component of volatility can be a more important driver of macroeconomic variables, and is thus a more suitable object of policy focus. Policymakers have spoken of the challenges of incorporating the effects of volatility /uncertainty into a macroeconomic model, and quantifying their effects in a forecast setting. Our model offers a simple way to address this challenge using market data.
Jeremy Chiu works in the Monetary Analysis Division at the Bank of England, Evarist Stoja works at the University of Bristol and is a Bank of England consultant and Richard Harris is a Professor at the University of Exeter.
If you want to get in touch, please email us at firstname.lastname@example.org or leave a comment below.
Comments will only appear once approved by a moderator, and are only published where a full name is supplied.
Bank Underground is a blog for Bank of England staff to share views that challenge – or support – prevailing policy orthodoxies. The views expressed here are those of the authors, and are not necessarily those of the Bank of England, or its policy committees.