Independent Thinking^®

Booms and Busts: A Brief History of Capital Spending Cycles

Bill Gates has a good line on change: “People,” he says, “often overestimate what will happen in the next two years and underestimate what will happen in 10.” We see evidence of that now, in the rush to exploit Artificial Intelligence (as described in CapEx: Too Much of a Good Thing?) and the related stock market volatility. If the past is any guide, the capital being spent today will eventually pay off – but the going may not be smooth.

CapEx mega-cycles are generally characterized by the invention of a new, disruptive technology that requires the buildout of massive infrastructure. They tend to revolve around major productivity-enhancing innovations, such as new forms of communication, new energy sources or generation methods, or transportation. Cheap and easy credit can pump investment – so can government backing through direct funding or regulatory support. While government support and debt funding can help a new industry get started, too much of both can often create a glut of capital, which leads to malinvestment.

The early automakers, the advent of electrical generation, and more recently, the smartphone, had prolonged growth trajectories. The industries that grew up around those new technologies all experienced consolidation (meaning many of the early companies failed), as well as continued innovation and capital spending. In the end, however, these advances in technology massively enhanced productivity without any significant economic hangover when their growth eventually slowed.

Other technological innovations were punctuated by overzealous capital spending booms followed by devastating financial crashes. Post-crash, industry growth and investment continued, albeit at a more reasonable pace. Steam locomotion was an early, transformative example. Invented in 1797 in England, the first rail lines in the United States were built in 1827 by the Baltimore and Ohio Railroad. The ability to transport goods and people quickly and safely over long distances changed the nature and trajectory of westward expansion in the country and supercharged national economic growth. In 1862, the Pacific Railway Act authorized the construction of the first transcontinental railroad, which ultimately linked California with the rest of the nation. This and the end of the Civil War in 1865 sparked even more growth, fueled by the combination of government support and debts. That ended in 1873 when Jay Cooke & Co., the bank that financed many railroad firms using high levels of debt, went bankrupt in September of that year. A market panic ensued, ending with bankruptcy of around one-quarter of the 364 railroad companies, and by 1876, 14% of the labor force out of work.

Still, railroads continued to expand after the crisis passed to peak in 1916 at over 254,000 miles. The growth of the railroads and continued improvements in track and locomotive technology spurred ever better connectivity and transportation, speeding the overall development and expansion of the country. The best investments were in the companies that benefited from this development, such as Sears, Roebuck & Company and Standard Oil, both of which leveraged railroads to build their companies to national scale.

Fast-forward to the telecom and internet boom of the 1990s and a similar story plays out at a much faster pace. The telecom CapEx cycle really took off in the middle of the decade as a massive new source of demand (the internet) combined with new advances in fiber-optic technology and the passage of the Telecommunications Act of 1996 (which allowed for more competition and new entrants).

This set off a wave of investment. Technology-focused capital spending between 1996 and 2000 increased by 75.6% cumulatively.¹ Many of these big spenders were new entrants into telecommunications infrastructure, spurred by deregulation and further encouraged by a massive growth opportunity. While the internet usage and use cases continued to grow for decades, by 2001 it was clear too much capital had been deployed too quickly. Problems were exacerbated by fraud and too much debt: Global Crossing, WorldCom and Enron all went bankrupt due either to fraud or too much debt, or in some cases, both. The subsequent stock market crash resulted in massive declines, over 70% for both the tech-heavy Nasdaq and the S&P 500 Communications Services Sector Index.² As with the railways, many of the companies that invested in the fiber build-out of the 1990s were not around to see it come to fruition. While AT&T, Verizon and Cisco are all still around, Amazon and Google, which used the internet and telecommunications network to build massive global businesses, turned out to be better investments.

AI is at least as likely to generate new industries and avenues for growth. But how long before today’s investments pay off? Is AI the next smartphone or will the cycle look more like the telecommunications industry? At current spending rates, Herculean growth assumptions must be achieved over relatively short periods. The biggest players know that, however, and they – Amazon, Google, Microsoft and others – have massive balance sheets and are taking on very little debt.

As investors, we are reasonably confident that these companies, among a few others, will be able at present to focus on AI’s potential over the next 10-plus years, and not just the next two, which mitigates the worst-case volatility for this cycle. But we are also cognizant that in past cycles, the best investments were often the beneficiaries of the innovation, not the innovators themselves. Sears was a better investment than Union Pacific in the first decades of the 1900s, just as Amazon was a better investment than Cisco in the first decades of the 2000s. We are focused on finding companies in healthcare, financial services and manufacturing that will be the primary beneficiaries of this CapEx cycle.

Brian Pollak is a Partner and Portfolio Manager at Evercore Wealth Management and the Chair of the Investment Policy Committee. He can be contacted at [email protected].

¹ Bureau of Economic Analysis.
² Peak to trough 03/09/2000 to 10/09/2002.