(Bloomberg Opinion) -- Market participants and academics have devoted a great deal time analyzing whether Bitcoin and other cryptocurrencies are investments or something else. A paper published this month by Yale University researchers took a look at the inverse question: how might traditional investments be affected by cryptocurrencies? Surprisingly, the authors find that obvious sectors to be disrupted by crypto—finance, wholesale and retail—are not affected but others have reason to fear or cheer crypto.
The authors overstate their case a bit when they write the results can be thought of as “indicating the potential winners and losers from the current and future development of blockchain.” Their case is based on correlation, and of course, correlation is not causation. It’s not plausible that blockchain or crypto had a significant causal impact on major industries from 2011 to 2018. Crypto was too small and obscure, and when it got big in 2017 its prices were clearly being driven by crypto-specific factors.
A better way to think about it is macro factors affect both crypto and certain industries. Statistical analysis alone can’t tell us what those factors are, but it can help describe alternative future scenarios. Looking across the hundreds of regressions in the Yale paper, a clear pattern emerges. Businesses that deliver goods to consumers—beer for example—or non-virtual services—dry cleaning—do well in the environments that favor cryptocurrencies. Businesses that sell to other businesses or government—metal fabricators—or deliver content to consumers—book publishing—do badly. (Full disclosure: I own Bitcoin and other cryptocurrencies.)
This makes intuitive sense to me. A technophile/libertarianish future could allow empowered consumers to use crypto to buy the things they want directly from producers and service providers, with profits squeezed from intermediaries and commoditized suppliers. In this world, digital freedom has triumphed over regulation, and businesses that depend on intellectual property are weak. This does not mean that authors and other content creators starve, just that they’ll have to connect directly with consumers—probably via a crypto solution—to get paid. Similarly there would still be plenty of capital investment and jobs supplying goods and services to non-consumer-facing businesses, but these would be more like public utilities operating at cost.
I can equally imagine a world in which the promise of cryptocurrencies is co-opted by traditional businesses. Many of the initial promises of crypto have already been fulfilled by fintech initiatives either selling to or run by established financial institutions. Other crypto potential might be blocked by technical issues, lack of consumer acceptance or regulation. Instead of profits depending on direct consumer satisfaction, they are determined by complex economic and legal arrangements by industry. Relatively high prices consumers pay for simple products like beer and dry cleaning leak off among many intermediaries and government entities before reaching the producer.
This is only a general impression from reviewing extensive regression results on the edge of statistical significance. For example, in one table, the authors report 30 regressions of Bitcoin prices on industries. They use a 5 percent significance level, meaning that even if there is no relation between Bitcoin and any industry we would expect to see 1.5 apparently significant results. The authors find four, which is only a weak signal that there is any relation at all. Moreover, I reran the results using daily and weekly data instead of monthly, and equal-weighted portfolios rather than value-weighted, and different time periods. The results change considerably. The values for Ethereum and Ripple differ from Bitcoin, and the results from China differ from the U.S. There are well-known problems with industry data: the categorizations are dated and both too broad; different types of businesses are included in one industry category and too narrow; many companies are in multiple industries. Crypto prices also suffer from data issues and, more important, a rapidly evolving market.
Due to the statistical issues, these scenarios are not certain enough for investment decision-making. There are other sections of the paper that are more relevant. On the other hand, the overall pattern seems persistent enough to inform serious thinking about the future of the economy with and without crypto. Although the data are only suggestive, the alternative is mainly unsupported opinion.