Tattle-tails: Gauging downside risks using option prices

Introduction

Over the course of a few weeks in March 2023, Silicon Valley Bank, Signature Bank and Credit Suisse⁠—three banks with combined assets worth US$1 trillion—failed. Authorities reacted swiftly to limit the stress of these failures on the global financial system. However, market participants were initially concerned that these individual failures could result in a deep financial crisis, similar to the 2008–09 global financial crisis, because of the integrated nature of the financial system. If a crisis were to emerge, it would have significant adverse effects on credit access, economic growth and inflation around the world and would likely result in central banks lowering their policy rates.

Reflecting the concerns of market participants, yields of government bonds fell sharply, both in Canada and other countries (Chart 1). The fall in government bond yields seemed to indicate that market participants saw a large risk of a crisis emerging in Canada.

Chart 1: Government bond yields in Canada and around the world declined sharply

At the same time, the value of banks and equities fell, but these declines were not nearly as dramatic (Chart 2). These relatively small declines seemed inconsistent with market participants seeing a large risk of a crisis emerging in Canada.

Chart 2: The stock prices of global banks declined through the stress period

We seek to determine the relative likelihood of two alternative scenarios that were consistent with these changes in asset prices:

• Did the drop in prices of financial assets suggest that investors saw increased downside risks to Canadian inflation, economic growth and policy rates?
• Did a spike in uncertainty during the crisis simply lead to investors having an increased desire to hold safe assets, without changing their expectations for Canadian inflation, economic growth and policy rates?

We use information from options markets in Canada to gauge the likelihood of these two scenarios during the period of global financial system stress. The prices of options reveal the probabilities of a range of positive or negative outcomes for different assets.1 We use the prices of options on exchange-traded funds (ETFs) to obtain the probability of future prices (the distribution) for Canadian short-term bonds, the S&P/TSX 60 Index (TSX) and Canadian bank stocks.2 We observe three key patterns during the stress period:

• a large increase in uncertainty and downside risks to short-term interest rates
• downside risks to the value of bank stocks that were far less significant than would have been consistent with a large risk of broad financial system stress in Canada
• minimal downside risks to the TSX, even during the peak of the crisis

Taken together, these three patterns suggest that the fall in bond yields in mid-March 2023 reflected an increase in how much investors desired to hold safe assets amid an increase in uncertainty, rather than rapidly changing expectations for Canadian economic growth, policy rates or inflation.

Mixed signals in Canadian fixed-income markets

Canadian short-term bonds were likely the asset most impacted by the stress because:

• their prices fluctuate with changes to monetary policy and economic conditions
• investors value them as safe assets during times of rising uncertainty and stress

To gauge the impact of the stress on the short-term bond market, we obtain the probability of various future prices (the distribution) for XSB, a short-term bond ETF. Because XSB primarily holds bonds with low risk of default, its price is highly sensitive to the range of potential outcomes for the Bank of Canada’s policy rate over the next two years. Chart 3 displays this distribution on three dates that summarize how the risk of lower policy rates changed during the financial system stress:

• March 6 is the pre-stress reference
• March 24 shows the peak level of stress as signalled by traditional asset prices
• March 31 shows the point when the stress had started easing, after US and Swiss authorities took swift actions to address the crisis

Chart 3: The financial system stress increased uncertainty and downside risks to Canadian interest rates

Chart 3 shows that between March 6 and 24:

• the mode of the distribution shifted lower as Government of Canada bond yields declined
• the variance of the distribution rose as uncertainty about the level of Canadian short-term bond yields increased significantly
• downside risks to Canadian short-term bond yields increased

By March 31, concerns about a severe shock to Canadian short-term bond yields had eased, as shown by the rightward shift of the mode and lower variance of the distribution in Chart 3. But the distribution did not fully recover to its March 6 level. This partial recovery suggests that market participants still expected that the financial system stress could have a moderate effect on access to credit in Canada, reducing the need for higher policy rates.3

This elevated uncertainty and increased downside risk to Canadian short-term bond yields suggests that market participants may have been concerned that the financial system stress would spread to Canada. The resulting impact on economic conditions could subsequently lead the Bank to substantially reduce policy rates. However, the uncertainty and reduced yields could also be consistent with a rise in both macroeconomic uncertainty and investors’ desire for safe assets at the peak of the financial system stress. To distinguish between these two plausible explanations, we expand our analysis to other asset classes. If the risk of a widespread stress event in Canada’s financial system was rising, the distributions of other asset classes would also have experienced rising downside risk.

Limited risk to Canadian financial institutions and economic growth

We apply the same methodology that we used for XSB to ETFs that follow stocks, namely the TSX index and the major Canadian banks.4 Ouellet Leblanc, Fontaine and Shotlander (2021) explain that the value of a stock can be determined by three factors:

• the rates at which investors discount safe cash flows back to the present
• the equity-risk premiums used to discount risky cash flows back to the present
• investors’ forecasts of future earnings

Assessing the impact of a shock on stocks involves analyzing the impact of the shock on these three factors and the relative magnitudes of these impacts.

The financial system stress likely affected the value of Canadian stocks through more than one of these factors. Additionally, some of these factors may have had offsetting effects on stock values during the stress period. For example, a decrease in credit availability would weaken demand, reducing future earnings. But weaker demand would lead to a decline in inflation. Lower inflation could lead to lower policy rates, which would result in lower risk-free rates. Finally, stock values would be affected by the changing risk premium that investors demand given the perceived riskiness of these future cash flows. If the magnitude of these offsetting effects is similar, the net impact on stocks would be negligible.

Chart 4: Declines in Canadian bank stocks and downside risks were not consistent with risks of a domestic banking sector stress

The global financial system stress led to a decline in the value of the stocks of major Canadian banks, as illustrated by the leftward shift of the distribution (Chart 4). This suggests that changes to expectations for future bank earnings or to the risk premium outweighed the positive impact of lower risk-free rates on the value of bank stocks. The variance of the distribution widened between March 6 and 24, indicating that investors became more uncertain about the future value of Canadian banks.

But this uncertainty lessened as concerns about stress in the financial sector eased. Importantly, the magnitude of the risks for Canadian banks remained small throughout the period. This suggests that investors thought that risks to the Canadian financial system were limited. By the end of March, the distribution had not yet recovered completely, which implies that investors still perceived some risks to the value or profitability of Canadian banks.5 However, the absence of large downside risks in the distribution suggests that investors did not expect one or more Canadian banks to come under significant financial stress.

Chart 5: The stress created minimal downside risks for broad Canadian economic growth

Finally, to measure the spillover effects of the global financial system stress to broader Canadian economic growth, we apply this methodology to the TSX index. The initial decline in the TSX index was much smaller than the decline in the valuations of Canadian banks (Chart 5), indicating that investors viewed the risks to total growth as somewhat minor even during the peak of the stress.

Additionally, the variance of the distribution remained relatively unchanged throughout the stress period. Because the stress likely increased the risk premium demanded by investors, the narrow distribution suggests that investors believed that a decrease in policy rates would offset any decline in future earnings and thus not significantly affect stock values. By the end of March, concerns of spillover effects to the Canadian economy had dissipated: the mode nearly fully recovered to its March 6 level, and the variance of the distribution was even lower than before the stress.

The differences in the risks that investors perceived to the value of the TSX index, Canadian banks and Canadian short-term bond yields provide a more nuanced understanding of the scenario market participants were most concerned about in March 2023. The financial system stress clearly created uncertainty in financial markets and increased investors’ desire for safe assets. But investors did not perceive a high likelihood of an adverse scenario for Canadian banks or the Canadian economy.

Appendix

We use options on three Canadian ETFs to assess the risks to the Canadian economy caused by global financial system stress in March 2023 because of their higher level of total activity compared with options on other Canadian assets. Data on total volume are available on the Montréal Exchange (MX) website. MX quotes will change either when the contract is traded or a new ask quote is provided to the exchange. The observed change in prices therefore reflects transactable values.

Table A-1: Description of exchange-traded funds

Endnotes

References

Breeden, D. T. and R. H. Litzenberger. 1978. “Prices of State Contingent Claims Implicit in Options Prices.” Journal of Business, 51 (4): 621–651.

Shimko, D. 1993. “Bounds of Probability.” Risk, 6 (4): 33–37.

Ouellet Leblanc, G., J.-S. Fontaine and R. Shotlander. 2021. “What Cured the TSX Equity Index After COVID‑19?” Bank of Canada Staff Analytical Note No. 2021-3.

Acknowledgements

We’d like to thank Gitanjali Kumar, Jean-Sébastien Fontaine, Virginie Traclet and Bruno Feunou for their continued support and encouragement throughout the process of writing this note. We’re also grateful to Nicole van de Wolfshaar and Jordan Press for their editorial assistance.