|
One question we’ve been discussing on the reporting desk lately is the difference between US president Donald Trump’s impact on the foreign exchange markets in his current term compared with his first. Take yourself back eight years and there are some parallels. For example, Trump’s frequent, unpredictable tweets about US trade policy with China in his first term caused several fluctuations in the US dollar/offshore Chinese renminbi exchange rate. There was also geopolitical turbulence arising from US foreign policy, such as the missile strike that killed Iranian military commander Qasem Soleimani at the beginning of 2020, sparking concerns of a wider conflict in the Middle East. But compare FX volatility during the two terms – taking the Covid-19 pandemic aside – and some stark differences become apparent. Before 2020, years of co-ordinated low interest rates across G10 markets and limited interest in carry trades led to record-low volatility, with significantly reduced leverage being deployed in FX markets. And despite a series of erratic tweets from Trump, volatility was structurally suppressed. Fast-forward to the present and market conditions differ significantly. The ‘liberation day’ tariff announcements and their subsequent impact on US assets have flipped long-established correlations on their heads. In addition, volatility materially blew out, impacting spreads and liquidity. Since then, the frequent nature of Trump’s unpredictable social media posts regarding tariffs, the Middle East, and even Fed chair Jerome Powell, have left exchange rates incredibly jittery when initially reacting to the headlines. When rumours that Trump would fire Powell begun to circle in mid-July, euro/US dollar spot jumped 1.5% in just 30 minutes. And when those rumours were quashed later that day, the exchange rate reversed all gains as if nothing had happened. This kind of intraday volatility is an increasing challenge that electronic FX market-makers will have to deal with for the next few years. Rather than reacting to headline news, they are reacting to market activity. During the Powell episode in July, some dealers saw volumes jump sixfold in that half-hour period. It’s an opportune time not only to prove their worth to clients in these events, but also to make profit in their spreads. This is where machine learning and large language models could become vital, scanning market data, news headlines, liquidity and top-of-book spreads to recalibrate pricing, ensuring they widen spreads when volatility spikes and also contract when it comes back down. Furthermore, whereas pricing algorithms can pre-determine spreads for fixed events like central bank meetings or data releases, banks may have to rely more on adaptive market-making tools if these unpredictable market episodes persist. These algorithms can detect intraday volatility on the fly and the adaptive market-making engine would kick in to adjust spreads accordingly. But not all tier-two and tier-three regional banks that have stepped into electronic market-making will possess these adaptive tools. The consequence could be liquidity providers stepping out of the market when volatility spikes – potentially making FX trading even more jumpy. So, while more certainty now exists around tariffs, e-FX desks may have to stay on their toes for the foreseeable future.
|
