Warren Buffett, perhaps the most sophisticated investor of all time, has historically eschewed technology stocks, precisely because he never had a way to value them. He once joked to a group of students that he would fail anyone who answered an exam question asking them to value an internet company. This mindset is unfortunate, because Buffet is a patient investor, and technology companies would benefit from his stewardship. So what if we could change that? What if Warren Buffett were to meet Paul Romer, the economist who won the Nobel Prize two months ago for his theory formally linking innovation to growth?
Romer’s theory applied to companies
To greatly simplify Romer’s theory, growth from innovation is driven by two factors: the level of research and development (R&D) spending, and the productivity of that spending, R&D productivity. While Romer’s theory is macro-economic, meaning it pertains to the entire economy, other research has extended it to the company level.
Generally, people focus on the spending piece of Romer’s theory, because of an explicit prediction from it, known as the “scale effects” prediction. The prediction is that growth should be proportional R&D—so if R&D doubles, so too should growth.
Given this prediction, we should be surprised by a “time tested finding” in Strategy&’s 2018 Global Innovation 1000: “There is no long-term correlation between the amount of money a company spends on its innovation efforts and its overall financial performance.” But this apparent contradiction, ignores the other element of Romer’s theory, R&D productivity. The scale effects prediction only holds if R&D productivity is the same across companies, and over time. What if it isn’t?
What is R&D productivity?
Before answering the question of what happens if R&D productivity isn’t the same, it helps to define it. R&D productivity, like any other form of productivity, is output per unit of input. In the case of labor, productivity it is typically measured as dollars of output per hour of labor. In the case of R&D however, productivity measurement is less straightforward, because current R&D has almost no impact on current revenues. Even if we compare tomorrow’s revenues to today’s R&D, capital and labor play a much bigger role in determining revenues, so we want to account for them. Most importantly, the role of R&D is growth, so we want to compare today’s R&D to changes in future revenue.
There is a measure of R&D productivity that does all this. The measure is called RQ (short for research quotient) as an allusion to individual IQ—it’s essentially how smart companies are. Just as high IQ individuals solve more problems per minute, high RQ companies solve more technical problems per dollar. The technical definition of RQ is the percentage change in revenue from a 1% change in R&D. This definition matches the most common approach in economics to measuring the returns to R&D for an industry or an economy. More importantly, RQ matches the full set of predictions from Romer’s theory when extended to the company-level: firms with higher RQ—those that are better at R&D, have higher growth, higher market value, and can profitably invest greater amounts of R&D.
Utilizing Romer’s theory for value investing
Not surprisingly, companies differ dramatically in their RQ. This explains why Romer’s scale effects prediction doesn’t hold in the Strategy& report. R&D spending will only explain growth when R&D productivity (RQ) is taken into account.
The figure below captures how important these RQ differences are to investors. The figure compares three portfolios: the S&P500 (market), a portfolio of the top fifty companies as ranked by their RQ (high RQ50), and an opposing portfolio of the bottom fifty companies as ranked by their RQ (low RQ50). Each begins with an initial investment of $1,000 in July 1981, and ends in December 2016. The figure shows that while investing in the S&P500 would generate a gain of 4100%, the high RQ50 would have generated a 7600% gain, almost double the returns. What may be more interesting however, is that the S&P500 outperforms the low RQ50 by an even greater degree. So conducting R&D in and of itself, is not a path to high returns. High returns require that R&D be in capable hands.
The figure is not intended to advocate investing in the RQ50 portfolio. Instead, it calls attention to an opportunity for value investing in technology, one created by marrying Warren Buffett (value investing) to Paul Romer (endogenous growth theory). Exploiting the opportunity requires understanding a company’s R&D productivity when other’s don’t. The returns of the high RQ50 only exceed the market because investors historically haven’t recognized a company’s R&D productivity until it’s been translated into a consistent level of earnings growth. Until that time, those stocks have been undervalued (or overvalued in the case of the lowRQ50).
This article originally appeared on Forbes