HSBC Chief Asia Economist Frederic Neumann said the current backdrop carries three signals that resemble the period before the 1997 Asian financial crisis: rising US Treasury yields, a roughly 57% depreciation in the Japanese yen over two years, and a technology boom now driven by artificial intelligence. But he said Asia’s vulnerability is no longer the same as it was in the late 1990s. The bigger risk now, in his view, is the region’s dependence on US demand for AI hardware.
US Treasury yields are the closest parallel to the pre-1997 setup
According to CNBC, Neumann identified US government bond yields as the most important similarity with the period leading up to the 1997 crisis. The 10-year US Treasury yield climbed from 5% in October 1993 to near 8% in November 1994. Even after easing to around 7% by April 1997, it was still about 200 basis points higher than four years earlier.
Against that comparison, the current cycle has taken the 10-year yield from a low of 0.5% in August 2020 to 4.79% earlier this week. Neumann said the rise has unfolded over six years, but this year alone it has jumped about 80 basis points from 3.9% in February.
Yields serve as a global pricing benchmark for capital. They shape borrowing costs, capital flows, and asset valuations. Neumann said the 1994 surge made dollar funding more expensive and pulled capital back toward the United States. Asian companies that had borrowed heavily in short-term dollar debt then struggled to roll that funding, setting off pressure that later fed into both currencies and debt markets.
He added that Asia’s external debt burden is much lighter now, but a higher discount rate can still compress tech valuations and raise the hurdle for investment. If US companies slow the pace of AI hardware spending, the effect could first show up in orders and then in Asian exports.
The yen is the second signal, with rare joint intervention in late July
Neumann’s second point of comparison is the yen. In 1995, the yen reached a cycle peak of 80 against the US dollar, then weakened to around 130 by April 1997. In the current cycle, it fell from about 103 in 2021 to 163 in July this year. It later recovered after rare joint action by Washington and Tokyo, and has recently traded around 160.
The report said that intervention was the first of its kind in a quarter century. Japan’s Ministry of Finance and the US Treasury entered the market together to buy yen on July 31. It was the first joint currency operation by the two governments since 2011, and the first coordinated move to support the yen since 1998, when the Japanese currency had fallen to its weakest level in about 40 years.
The third similarity is sentiment: internet optimism then, AI enthusiasm now
Neumann also pointed to market psychology. Before the 1997 crisis, the commercialization of the internet fed a broad wave of optimism around technology. He said today’s AI boom is creating a similar sentiment structure.
That does not mean the two periods will end the same way, but he sees the combination of tech optimism and changing funding costs as a setup worth watching.
Neumann says the differences between 1997 and 2026 are larger than the similarities
Neumann said plainly that the gap between 1997 and 2026 is bigger than the overlap. The main difference is the direction of capital flows. In the 1990s, most Asian economies were capital importers, meaning they absorbed more foreign investment than they invested abroad, while domestic savings were not enough to fund spending commitments.
He said that structure made the region highly sensitive to rising dollar funding costs and to investor unease around the yen. Those forces became key triggers for regional stress. Today, by contrast, Asia has become a capital exporter. Under that structure, higher US funding costs and a weaker yen are no longer the main pressure points.
The more immediate risk is Asia’s dependence on US AI hardware demand
Neumann said the change in financial structure does not make Asia immune. The more direct issue for the region is its exposure to the US AI hardware boom, which is currently supporting growth across several Asian economies.
He specifically pointed to exports of AI-related electronics as an important growth pillar for South Korea, Japan, Taiwan, and Singapore. If higher US Treasury yields and funding costs eventually curb AI hardware investment, or if yen volatility disrupts global capital markets, demand for regional exports could turn lower and economic momentum could cool with it.