Picture a Canadian factory boss in the early 2000s. A supplier in China offers the same components her firm has been buying domestically, but at a fraction of the price. The decision looks easy. Lower costs mean fatter margins, more competitive prices, and room to grow. But there is a quieter question lurking underneath: will cheaper parts make her company more inventive, or less?
For years, the common assumption among economists was that buying cheap inputs from low-wage countries should give innovation a boost. A new analysis distributed as an NBER working paper reaches the opposite conclusion. Studying thousands of Canadian manufacturers, the authors find that importing components from low-wage countries was linked to a decline in research and development spending.
Two forces pulling in opposite directions
The study was conducted by Wulong Gu of Statistics Canada, Alla Lileeva of York University, and Daniel Trefler of the University of Toronto’s Rotman School of Management. Before laying out their data, they describe two opposing pressures that offshoring places on a company’s incentive to innovate.
The first pressure is positive, and it has a long history in economics. When a firm buys cheaper parts, its costs fall and it can sell more. Because innovation makes every unit of output a little more profitable, selling more units raises the payoff from inventing something new. The authors call this the “innovation-offshoring complementarity.” It tends to matter most for large firms whose new products can capture a lot of additional sales.
The second pressure works the other way, and it is the heart of the paper. Companies usually sell an older generation of a product alongside whatever newer version they are developing. Older products tend to rely more heavily on cheap, lower-quality components, while next-generation products lean on higher-quality inputs. When the price of cheap Chinese components drops, it lowers the cost of making the old product more than it lowers the cost of making the new one. That shrinks the financial gap between the two, and it is that gap that motivates a firm to push toward the next generation. The authors call this the “relative-cost channel.” When it dominates, the incentive to innovate weakens.
An intermediate input, in plain terms, is a part or component a company buys to build its own product, rather than a finished good it resells. A semiconductor inside a phone is an intermediate input; the phone is the final good. The authors point to engineering case studies, including one on semiconductors, where cheap offshored components took so much cost out of older designs that developing newer ones became less attractive.
Tracking 6,024 firms across a decade
The researchers assembled a balanced panel of 6,024 Canadian manufacturing firms covering 2002 through 2011, drawing on six Statistics Canada datasets. A feature of their data is unusually detailed: for each firm they could see, at a fine product level, both what the company produced and what it imported. That let them separate true intermediate inputs (goods the firm bought but did not make itself) from finished goods it might import for resale.
They then measured how each firm’s imports of intermediate inputs from low-wage countries, mostly China and Mexico, changed over the decade, and compared that with changes in the firm’s R&D. The authors use a broad definition of R&D that includes a firm’s own research, research it contracted out to others, and payments to acquire technology such as patents and licenses.
A simple comparison would run into a familiar problem. Firms that import more might differ from other firms in hidden ways that also affect their research choices, making it hard to separate cause from coincidence. To address this, the authors built what economists call a shift-share instrument. The idea is to isolate a part of the change in offshoring that comes from an outside force unrelated to any individual firm’s circumstances. Their outside force was the often-dramatic improvement in the quality of Chinese-made components over the decade, measured using United States import data so that it would not be tangled up with conditions inside Canada.
By their estimate, the quality of Chinese intermediate goods relative to high-wage suppliers rose sharply during this period, with the median quality of Chinese products relative to Canadian ones rising by roughly 200 percent. They treat these quality jumps as roughly random across products, and run a series of tests to support that assumption.
What the analysis found
The results point in one direction. The authors estimate that the rise in offshoring of intermediates from low-wage countries reduced R&D spending by about 15 percent over 2002 to 2011. That total splits into two roughly equal pieces. Firms that were already doing research cut back on it, and firms that were not doing research were discouraged from starting. Extending the trend out to 2022, they project a 28 percent reduction.
Offshoring was also linked to a 7 percent drop in employment at these firms, consistent with the idea that domestic labor was being replaced, in part, by the foreign labor embedded in imported components.
One finding helps the authors distinguish their explanation from competing ones. Much of the existing research on Chinese imports treats them as a competitive threat that eats into a firm’s sales and profits, a “demand shock.” If that were driving the results here, profits should have fallen. Instead, the authors find that profits rose. They interpret rising quality of Chinese components as a “supply shock,” a free reduction in costs that lifts profits for the firms doing the importing. The negative effect on innovation, in their reading, comes from the relative-cost channel rather than from competition squeezing firms.
Big firms versus everyone else
The model predicts that the positive, sales-driven channel should win out only for firms whose innovations meaningfully expand demand, which the authors associate with multinational enterprises. The smaller, more incremental innovations of non-multinationals should be governed by the cost channel.
That is what the data suggest. When the researchers split their sample, all of the estimated decline in innovation came from non-multinationals, while multinationals showed no such drop and may have increased research. The authors caution that their estimates for multinationals carry large statistical uncertainty. They also find direct evidence of the substitution at the core of their story: as Chinese quality rose, firms shifted their sourcing away from United States and other high-wage suppliers toward Chinese ones, which accounts for much of the supply-chain restructuring seen in Canada over the decade.
What to keep in mind
A few caveats sit alongside these findings. The study covers Canadian manufacturers, whose deep integration into United States supply chains may make them unusual; the authors conjecture, but do not prove, that results would look similar in the United States. Because they study a fixed set of firms that survived the whole period, they could not examine companies that entered or exited, though they argue that including those would likely make the negative effect on innovation larger, not smaller.
The authors are also describing a particular kind of offshoring, the importing of low-quality components, rather than offshoring in general. Their results do not say that buying parts from technologically advanced economies harms innovation; in fact, offshoring from high-wage countries showed no negative effect in their tests. The takeaway they emphasize is narrower and pointed: when firms in a rich economy lean on cheap, lower-quality inputs from low-wage suppliers, the move that lowers costs today can also lower the reward for inventing tomorrow.
