REF: 2021

Impact case study

New approaches to forecasting financial markets


[Martin’s work is furthering] our understanding of the link between asset prices and macroeconomic fundamentals, and helping us implement practical tools [which] can allow us to make up-to-date asset price assessments for financial and monetary stability purposes.

Gertjan Vlieghe

Member of the Bank of England’s Monetary Policy Committee

Professor Ian Martin

Research by

Professor Ian Martin

Department of Finance

Research at LSE has produced innovative ways to forecast the behaviour of assets in exchange markets, improving how financial institutions can analyse volatility and market “bubbles”. 

What was the problem? 

The equity premium – the difference between expected returns on risky investments and the certain return on risk-free securities or portfolios – is one of the central quantities of finance and macroeconomics. This is not directly observable but can be inferred using observed stock prices and other market variables.  

Institutions such as central banks, monetary policymakers, and asset management companies all need ways to forecast the behaviour of financial asset markets to calculate this risk premium. 

Many existing forecasts based on, say, macroeconomic or accounting data, can only be exploited by investors or policymakers with a lag, not in real time. For example, it is notoriously difficult to forecast exchange rates. Typically, they are estimated using historical data, which is an imperfect measure.  

Designing a more accurate appraisal of forward-looking risk premiums can therefore improve the performances of investments. It can also help with understanding whether a market is “bubbly” and experiencing a destabilising rise in prices. If central banks and other institutions can spot such market bubbles, they can improve their efforts to maintain stable financial systems.  

What did we do? 

Professor Ian Martin’s research, published in a series of papers between 2013 and 2020, has introduced new methods for forecasting the behaviour of stock markets and currency markets. The distinctive feature of his work is its reliance only on observable asset prices, making the resulting forecasts observable in real time.   

Firstly, Martin defined a new index of volatility, the SVIX index, which represents the market's expectation of 30-day forward-looking volatility. Derived from the price inputs of the S&P 500 index options, Martin’s SVIX index provides a measure of both market risk and investors’ sentiments. It can be used to measure the expected return on the market – the so-called equity premium. Using historical option prices, Martin constructed the index for the period 1996 to 2012, which showed that the equity premium is (even) more volatile than had previously been understood.  

Subsequent research with Professor Christian Wagner (WU Vienna University of Economics and Business) built on this insight to derive a formula for the expected return on individual stocks, based on observable option prices. This outperforms other standard measures empirically and makes distinctive predictions that challenge conventional wisdom by showing that risk premiums are extremely volatile across stocks, as well as over time.  

In parallel work conducted with PhD student Lukas Kremens between 2016 and 2018, Martin made new theoretical and empirical contributions to the classic – and notoriously difficult – challenge of forecasting exchange rates. Kremens and Martin’s theory improves on existing work by not making the empirically implausible assumption that investors are risk neutral. The resulting measure was shown to work very well empirically, outperforming the standard competitors in forecasting differential currency movements.  

Research conducted in 2018/19 with Can Gao, a PhD student at Imperial College London, provides a real-time, forward-looking measure of “bubbliness” in stock markets. The indicator, which is based on option prices, valuation ratios, and interest rates, was unusually high during the late 1990s, suggesting that investors’ dividend growth expectations were unreasonably optimistic at the time, shortly before the dot-com bubble burst. This suggests that Gao and Martin’s measure can help to reveal irrational beliefs about fundamentals, and so is an indicator of measures associated with financial fragility.   

The most important conclusion reached from this body of research is that risk premiums are substantially more volatile, both over time and across stocks, than the literature has previously acknowledged, and that real-time measures like these can improve our estimation of them.  

What happened? 

This research has generated new debate about ways to forecast financial assets and improved how financial institutions understand risk premiums. It has also directly shaped the work of both public and private sector financial institutions. 

Professor Martin’s work has, for example, provided central banks with new tools to support the notoriously difficult task of forecasting exchange rates. In 2017, Martin presented his novel approach to the Bank of England, and he was asked to provide the data to compute the currency appreciation measure he had proposed, so that such forecasts could be submitted to its Monetary Policy Committee (MPC). They have since made use of it in formulating their views on exchange rates risk premiums. 

The real-time, forward-looking measure of “bubbliness” in stock markets is also proving useful for monetary policymakers at the Bank of England, where Martin served as an academic consultant in 2018/19. His indicator is, according to a senior adviser there, particularly helpful to the work of the Bank’s Financial Stability Strategy and Risk (FSSR) Directorate, as it can provide them with bubble signals with “a high level of confidence” and in real time, helping to inform the Bank’s asset valuations, and assessments of risk. 

His work has also been presented to the Federal Reserve Bank of New York (the Fed), the Bank of Japan, and the European Central Bank, which implemented several versions of the indicators developed to detect changes in market perceptions. By supporting the development of more robust monetary policy, and helping central banks to spot destabilising market “bubbles”, the work aids efforts to maintain stable financial systems in the UK and elsewhere.  

Some of the world’s biggest private financial organisations also recognise the research’s significance. In 2017, quantitative investment giant AQR Capital Management named the work with Wagner on the expected return on individual stocks as a finalist for its Insight Award.  

It has also been influential within BlackRock – the world’s largest asset manager. Former BlackRock senior portfolio manager, Dr Brandon Bates, noted the work: “teaches us how to think about markets, how to impose structure on a problem with formidable complexity”. For these institutions, the interest is in the implications for understanding equity risk premiums, and for informing asset allocation across classes. Martin’s measure allowed them, according to Bates, “a much richer opportunity set for tactical asset allocation than anyone at BlackRock previously thought.” He also noted that Martin’s work had been influential in the context of foreign currency exposure, and on setting the parameters for modelling the timing of trades in different market conditions. 

At the scale at which BlackRock and others operate, the influence of Martin’s work has very far-reaching effects on investment returns. By helping investors take better-informed risks, with a more efficient allocation of capital, the macroeconomic benefits are substantial.