Systemic risk and the macroeconomy: An empirical evaluation

Stefano Giglio, Bryan Kelly, Seth Pruitt

03 April 2015



The Global Crisis of 2007-09 has made systemic risk a focal point of research and policy, and has established the financial sector as its centre of analysis. The empirical side of the literature focuses on measuring distress in financial markets. This has produced a staggering variety of systemic risk proxies, many hoping to serve as an early warning signal of market dislocations like those observed during the Crisis. Bisias et al. (2012) provide an important qualitative survey of more than thirty such measures.  This variety reflects a “lack of specificity,” that according to Nobel laureate Lars Hansen, “could undermine the assessment of alternative policies.”

Systematic risk’s usefulness for policymakers

So a simple initial question is: Do any of the measures contain predictive information that is useful to policymakers? We investigate this question in our recent paper (Giglio et al. 2015). If policymakers or regulators plan to police systemic risk, they must do so out of a concern that it impacts welfare. That is, if systemic risk were solely a feature that affects asset markets without metastasising into the consumption or production of households, there is no obvious reason why it needs be regulated.1

Thus, our starting point is to argue that the policy-relevance of a systemic risk indicator depends on its informativeness regarding how financial distress translates into real macroeconomic outcomes. Indeed, the ECB (2010) defines systemic risk as a risk of financial distress “so widespread that it impairs the functioning of a financial system to the point where economic growth and welfare suffer substantially.” We provide a first attempt to broadly quantify the extent to which the systemic risk literature as a whole can be used to predict bad macroeconomic times on the horizon.

Systemic risk and macroeconomics shocks

We find that very few of systemic risk measures help forecast bad times. However, we also find it is helpful to combine the measures into an index, and we propose a new way of doing so. Our systemic risk index demonstrates highly significant forecasting power for recessions. During normal times, we estimate there is a 1-in-5 chance of quarterly IP growth being as low as -1.4 percentage points (annualised). But during several historical episodes (including the recent Global Crisis), the systemic risk index predicted a 1-in-5 chance of quarterly IP growth dropping by at least three percentage points (annualised).

Interestingly, the relationship between systemic risk and future macroeconomic shocks is not symmetric. IP shocks’ median, or even their probability of being largely positive, is little affected by systemic risk. The strong relationship is between systemic risk and the probability of bad macroeconomic shocks.

In fact, this empirical finding is quite in line with theoretical literature like Bernanke and Gertler (1989), Kiyotaki and Moore (1997), Bernanke et al. (1999), Brunnermeier and Sannikov (2010), Gertler and Kiyotaki (2010), Mendoza (2010), and He and Krishnamurthy (2012). These theories predict that distress in the financial system can amplify adverse fundamental shocks and result in severe downturns or crises. However, they also imply that the absence of stress is not necessarily a trigger for good macroeconomic times ahead. That broad pattern is exactly what we find in the data.

A set of new stylised facts emerges from our investigation.

  • First, we find that a select few systemic risk measures tell us about future macroeconomic shocks. 

Measures of financial sector equity volatility perform well in a variety of specifications. This result is optimistic about the feasibility of constructing systemic risk measures for a variety of markets, because equity prices are perhaps the most widely available financial market data worldwide.

  • Second, while financial sector equity volatility is quite informative about future bad times, equity volatility in the nonfinancial sector appears to have little, if any, predictive power.

Suppose that we, following a literature including Bloom (2009), take volatility to represent uncertainty. This suggests an important distinction between uncertainty in the financial sector and uncertainty in other industries. Uncertainty in the financial sector predicts an increased chance of bad macroeconomic times ahead. Perhaps this distinction between financial and nonfinancial equity price volatility could lead to economic insights into the nature of systemic risk.

Finally, we find that a rise in systemic risk predicts an increased probability of a large drop in the Federal Funds rate. 

This suggests that the Federal Reserve takes preventive action amid elevated risk levels. However, it appears historically as though such preventative action failed to fully counteract the risk of economic downturns.

Concluding remarks

Our analysis says that notable value has been added by the burgeoning literature on systemic risk measurement. In particular, several measures give us important information about the probability of future macroeconomic shocks, and combining them together helps. Nonetheless, the results distinguish the more promising avenues for future research to take, of which we highlight two. 

  • First, models connecting systemic risk to macroeconomic outcomes will require solutions methods allowing for asymmetric responses. 

This point, made theoretically by Brunnermeier and Sannikov (2012), finds strong support in the data we analyse. 

  • Second, it will be important to better understand the different forces that drive valuations for financial versus nonfinancial firms.  

In the least, this will direct policymakers’ attention to relevant price swings, and potentially it could suggest optimal policies aimed at stemming the rise of systemic risk in the first place.


Bernanke, B, and M Gertler (1989), “Agency Costs, Net Worth, and Business Fluctuations,” The American Economic Review, 79(1), 14–31.

Bernanke, B S, M Gertler, and S Gilchrist (1999), “The Financial Accelerator in a Quantitative Business Cycle Framework”, Handbook of Macroeconomics, 1.

Bisias, D, M Flood, A W Lo, and S Valavanis (2012), “A Survey of Systemic Risk Analytics”, Working Paper, Office of Financial Research.

Bloom, N (2009), “The impact of uncertainty shocks”, Econometrica, 77, 623–685.

Brunnermeier, M K, and Y Sannikov (Forthcoming), “A Macroeconomic Model with a Financial Sector”, The American Economic Review,

European Central Bank (ECB), 2010, “Financial networks and financial stability,” Financial Stability Review, 2010, 155–160.

Gertler, M, and N Kiyotaki (2010), “Financial intermediation and credit policy in business cycle analysis”, Handbook of Monetary Economics, 3, 547.

Giglio, S, B Kelly and S Pruitt (2015), “Systemic Risk and the Macroeconomy: An Empirical Evaluation”, Working Paper, Chicago Booth.

Hansen, L (2013), “Challenges in Identifying and Measuring Systemic Risk”, Working Paper, Chicago Booth.

He, Z, and A Krishnamurthy (2012), “A Macroeconomic Framework for Quantifying Systemic Risk”, Working Paper, Chicago Booth.

Kiyotaki, N, and J Moore (1997), “Credit Cycles”, Journal of Political Economy, 105(2), 211–248.

Mendoza, E (2010), “Sudden stops, financial crises, and leverage”, The American Economic Review, 100(5), 1941–1966.


[1] A systemic risk measure may be informative about many things, such as the valuation of derivatives or other risky assets. But unless this ultimately translates into welfare of households, the argument in favour of regulation is unclear.



Topics:  Macroeconomic policy

Tags:  systemic risk, macroeconomic shocks

Assistant Professor of Finance at Booth School of Business, University of Chicago

Associate Professor of Finance, University of Chicago Booth School of Business

Assistant Professor of Finance, WP Carey