Innovation is the foundation of growth. As innovation spreads to the physical industries, the result is higher wages, lower costs, and a more dynamic economy, as we showed in our recent report, The Coming Productivity Boom.

But innovation in physical industries  has proceeded slower than we might have wanted, In part, that’s because promising technologies are being hindered at the state and local  level. Consider the humble eye refraction. Some companies are rolling out technologies that enable ordinary people to do eye refractions remotely, using computers or smartphones. Such a technology has the potential to greatly reduce the cost of updating and renewing contact lens prescriptions, while enabling people to check their eyes more often.

Unfortunately, in Connecticut, the state legislature is now considering a bill that would undercut the use of remote eye refractions. The bill says, in part:

No provider shall issue an initial prescription to or renew an initial prescription for a patient without having performed an in-person evaluation and an eye examination of the patient.

In other words, everyone has to make an expensive and time-consuming in-person trip to an eye doctor to get a prescription renewal for contact lens, even if the remote refraction says no change. This requirement could be especially costly in Connecticut, where optometrists make $192,870 per year on average, the highest in the country, according to the Bureau of Labor Statistics.* Indeed, Connecticut is the only state where optometrists make more than family and general practitioners, based on BLS data.

Remote refraction can and should never replace in-person visits to optometrists and ophthalmologists. But in a world where health care costs are increasingly squeezed, it seems silly and downright embarrassing for a forward-looking state like Connecticut to inhibit a technology that could make people better off at a lower cost.

*These figures are based on the May 2016 Occupational Employment Statistics from the BLS. See this page for a list of top-paying states for optometrists. This data does not cover self-employed workers.

According to PPI estimates, rising labor costs accounted for almost $65 billion in added health care costs in 2015, or 47 percent of the total increase in personal health care spending (as reported by the latest projections from the actuaries at the Centers for Medicare and Medicaid). By contrast, IMS reports that net spending on prescription drugs rose by only $24 billion in 2015, or 18 percent of the rise in personal health care spending.

This result, which updates our previously published data for 2014, fights the prevailing narrative that healthcare spending is primarily driven by rising drug prices.. Instead, the big increase in labor compensation in 2015 is mainly being impelled by the rapid growth of health care employment.  The number of workers in the nation’s hospitals, physician offices, and nursing homes increased by 429,000 in 2015, compared to an annual average of 226,000 for the previous 5 years.

Unfortunately, this trend of rapidly rising healthcare employment driving overall spending appears to have accelerated in 2016. Healthcare employment in the 12 months ending July 2016 was more than 500,000 above its year earlier level, suggesting that healthcare labor costs will surge even more in 2016.

Meanwhile data from the BEA seem to show the pace of pharmaceutical spending slowing in 2016. Consumer spending on prescription drugs in August 2016 was 4.3% over a year earlier, less than half as fast as in 2015. There’s no guarantee that this slower pace will continue for the rest of 2016, but so far it’s a good sign.

I am speaking Thursday at a Cato conference on The Future of U.S. Economic Growth, with a politically diverse group of speakers including Martin Baily, Robert Gordon, Brad DeLong and Erik Brynjolfsson. My panel is entitled”What is to be done?,” and focuses on feasible policy solutions.

In preparation for the conference, I put together an essay on “Hacking the Regulatory State.” Part of the essay covers the need for a Regulatory Improvement Commission, but I also laid out some ways that the FDA can be reformed to speed up economic growth. Here’s an excerpt from the essay:

2. Approval Criteria at the FDA

The FDA is one of the fastest growing agencies in the federal government. In 2000, the FDA employed 12 workers for every 1,000 in the pharmaceutical, biotech, and medical equipment industries.  Now the FDA employs 18 workers for every 1000 private-sector pharmaceutical, biotech, and medtech workers.

Not surprisingly, the intensity of FDA regulation has also increased by 40 percent since 2000, according to a recent paper from the Progressive Policy Institute (Carew, 2014). That’s based on a new measure of regulatory intensity that applies a semantic analysis of written rules, looking for such restrictive words as “shall” and “must” (Al-Ubaydli and McLaughlin, 2014).

The same period has also been notable for an extraordinary amount of public and private spending on biosciences R&D. In 2012, for example, U.S. industry, government, and academic institutions spent roughly $100 billion on biosciences-related research and development, second only to the roughly $125 billion invested in computer and information sciences-related R&D. In recent years biosciences R&D has averaged somewhere between one-third and one-quarter of total civilian R&D.

This R&D spending has propelled tremendous scientific advances over this stretch. Yet so far, too few of these scientific advances have been translated into usable innovation. This problem is well-accepted. NIH set up a new National Center for Advancing Translational Science in fiscal year 2012, specifically to “develop innovations to reduce, remove or bypass costly and time-consuming bottlenecks in the translational research pipeline in an effort to speed the delivery of new drugs, diagnostics and medical devices to patients.”

I will argue here that accelerating commercial innovation in biosciences requires “recoding” the criteria by which the FDA approves new drugs and devices. In particular, the sole focus on “safety and efficacy” has the effect of almost guaranteeing that potential disruptive innovations are not approved. What’s more, the pharmaceutical and device companies have a deep understanding of the FDA’s approval process, and therefore they do not pursue such disruptive innovations. Similarly, venture capitalists shy away from funding innovations that are not approvable.

Here I’m using the disruptive innovation in the classic Clayton Christensen sense — a product or service that starts out with somewhat worse performance than what’s on the market right now, but much better economic or other characteristics. So when mobile phones originally were being widely sold, the quality of calls was lower than using wired handsets. Similarly, the early personal computers were far less powerful than mainframes or minis.

The problem is that the FDA interprets the “safety and efficacy” standard as meaning at least as safe and clinically efficacious as anything on the market currently. That immediately rules out an innovation that is safe, much cheaper, but not as efficacious as best medical practice. So if the FDA had been in charge of the phone or computer markets at the time, early mobile phones and personal computers would have not been approved for sale because they provided inferior quality to existing products.

As a result, the FDA approval criteria systematically screen out disruptive innovations. What’s more, the pharmaceutical and device companies, and even the venture capitalist supporting start-ups, are all too aware of the FDA’s decision-making process,  and are therefore unwilling to fund potential disruptive innovations.

What’s the solution? First, don’t weaken the safety requirement at all. The FDA is a key guardian against harmful products.

Second, separate the efficacy requirement into two parts — clinical efficacy, and economic efficacy. Allow innovating companies to present evidence that their potential new product reduces the amount of labor and other resources needed by the healthcare system, as compared to existing products or treatments. A new product needs to show both clinical efficacy and economic efficacy, but needs to be superior to existing products on just one of those measures.

Such a broadening of the FDA approval criteria won’t be easy to put into place, but could have enormous impacts on the incentives for research and development. If we want medical innovation and lower costs, we need to change the rules of the game.