The Global Macro Key to Determining Domestic Tail Risks
- By Sam van de Schootbrugge
- 5 min read
Summary
- A new University of Cambridge Janeway Institute working paper is the first to examine how the tails, rather than the mean, of domestic growth are influenced by global macro factors.
- It finds foreign vulnerabilities, such as equity volatility, credit-to-GDP growth and global financial conditions, are key drivers of domestic macroeconomic tail risks.
- Developments in foreign countries have a significantly larger effect on domestic tail events than domestic developments do.
Introduction
Disaster risk, or low-probability left-tail events, may at least partly explain excess returns in several markets. Essentially, investors price in the very unlikely event of a large negative cash flow realisation (or a change in their discount factor) to compensate for the risk of big financial losses. In FX markets, the ‘peso problem’ comes close to explaining returns from carry trade strategies.
The problem with pricing disaster risk, however, is that predicting unlikely events ex-ante is hard. And so, a new University of Cambridge Janeway Institute working paper uses global macroeconomic data to document how crucial foreign vulnerabilities are in determining the likelihood of economic growth ‘disasters’. They find:
- Higher foreign equity volatility is associated with significant reductions in the left tail of domestic GDP in the near term (less than one year).
- Faster global credit-to-GDP growth and tighter global financial conditions exert a significant negative influence on the left tail of the GDP-growth distribution.
- These foreign indicators can provide advanced warning of crisis episodes, with foreign shocks explaining around 71% of the variation in domestic GDP at the fifth percentile of the distribution.
International Spillovers
The literature has identified several channels through which foreign developments affect domestic GDP growth. First, tighter global financial conditions can instantly impact domestic funding costs and risky asset prices, which is seen through the strong cross-country co-movement in asset prices. This, in turn, affects the distribution of future GDP growth.
Second, with financial institutions increasingly holding foreign claims, excessive credit growth and risk-taking abroad can generate losses for domestic financial institutions and cause spillovers to the wider economy. Third, a build-up in foreign vulnerabilities that triggers a downturn abroad can spill over to the domestic economy through broader macroeconomic channels – for instance, by lowering demand for domestic exports.
Quantile Regressions and GDP-at-Risk
The authors use a quantile regression setup in a global VAR model. By using growth quantiles, the paper examines the influence of foreign factors on the tails, rather than the mean, of domestic GDP growth. To account for foreign vulnerabilities, the authors use a weighted average of indicators in the rest of the world using bilateral-exposure weights from the IMF (trade) and BIS (financial), which allows them to capture differences among countries in terms of trade and financial linkages.
They estimate the GDP growth distributions of 14 advanced economies and define the bottom fifth percentile of the distribution as the GDP-at-Risk. According to former Governor of the Bank of England Mark Carney, GDP-at-Risk is a now widely used concept in financial stability monitoring and cost-benefit analysis informing macroprudential policy. And, intuitively, the authors want to know what causes changes in the GDP-at-Risk over time.
Research so far has identified financial market conditions, the evolution of credit, and real GDP growth as potential domestic predictors of domestic crisis episodes (i.e., they affect the GDP-at-Risk). And so, the authors use (i) realised volatility of equity prices, (ii) the three-year percentage point change in the aggregate private nonfinancial credit-to-GDP ratio, and (iii) the lagged one-quarter growth of real GDP as domestic covariates that capture movements in tail risk.
Uniquely, the authors also include the above three variables as foreign covariates. Essentially, they are testing whether changes in foreign financial and credit market conditions can also affect domestic GDP tail risks. Indeed, they find that higher equity volatility and credit-to-GDP in foreign countries is strongly associated with declines in domestic GDP (Chart 1). For example, in the first quarter, a one std. increase in foreign equity volatility is linked with a 0.7pp fall in the GDP-at-Risk.
The authors even find that the foreign financial risk factors more substantially affect domestic macroeconomic tail risks than domestic financial markets! Also, foreign credit growth has a significantly negative effect on domestic growth, but the size of the effect on the domestic macroeconomic risk outlook is similar to the domestic credit-to-GDP impact.
Greater foreign equity volatility affects domestic GDP growth across all quantiles of the distribution. In other words, more asset price volatility globally is no more likely to cause a severe domestic economic downturn than an average or very mild recession. This is not so for worsening global credit conditions, which makes it more likely a severe domestic downturn will follow.
Using Foreign Variables to Predict Crises
The authors use the information from foreign equity volatility, credit-to-GDP growth, and real GDP growth to predict UK GDP tail events from 1980 to 2020. They find two noteworthy observations. First, including these foreign variables improves the predictions of large recessions. Second, the impact of their inclusion on estimates of GDP tail events increases the further out the forecasting period goes, peaking at roughly the four-year horizon. In other words, the model is best at predicting tail events four years ahead.
Next, the authors examine whether there are any unique changes in the skew and kurtosis of the UK GDP distribution ahead of the Global Financial Crisis. In the run-up to the GFC (2004-2008), the model predicts a notable fall in the skew and a rise in the kurtosis of UK GDP growth – with the distribution becoming most left-skewed and fat-tailed around 2006Q2. This means the model provides a useful three-year advanced warning, not just of a leftward shift in the distribution of future GDP growth, but also a significant fattening in the left-hand tail, well ahead of the GFC.
The authors find similar results for all countries in the panel. For Germany, only with the introduction of foreign equity volatility and credit-to-GDP does the model predict a rise in downside risk around the GFC. According to the authors, this is clear evidence that when assessing German GDP tail risks, including global spillover effects is essential. For the US, the authors find that the skewness of 12-quarter-ahead US GDP growth shows a steady decline well ahead of the three US recessions in the sample.
How Do Foreign Shocks Affect Domestic Tail Risks?
As a final step, the authors decompose changes in domestic tail risks into foreign and domestic shocks. In the UK, for example, they find that three-year ahead tail risks increased in the run-up to the 1990-1991 recessions predominantly due to domestic drivers (red bars), particularly in domestic credit-to-GDP (Chart 3). Before the GFC, tail risks again increased significantly (bars declined), but this time driven largely by a build-up in foreign-weighted credit-to-GDP. This accords with the view that the GFC originated due to greater global interconnectedness and the rise in credit.
The authors quantify the importance of foreign conditions on domestic GDP. Across all 14 countries in the sample, foreign developments can explain 89% of the variation in tail risks over the next quarter. Over a three-year horizon, domestic factors play a larger role in domestic tail event probabilities, but still, foreign conditions can explain 71% of the variation.
Bottom Line
That greater financial integration has influenced domestic market conditions is well documented. However, few have investigated the origins of GDP-growth tail events. The evidence shows developments in trading and financial partner countries significantly govern the near- to medium-term downside risks to domestic growth. So, when searching for ‘steam-roller’ events with potentially large cashflow implications, consider the neighbours of the countries you are interested in – they are more likely to give you a better prediction of future crises.
Citation
Lloyd, S., Et Al., (2021), Foreign Vulnerabilities, Domestic Risks: The Global Drivers of GDP-at-Risk, Cambridge INET, https://www.inet.econ.cam.ac.uk/research-papers/jiwp-abstracts?wp=2102
Sam van de Schootbrugge is a Macro Research Analyst at Macro Hive, currently completing his PhD in international finance. He has a master’s degree in economic research from the University of Cambridge and has worked in research roles for over 3 years in both the public and private sector.