By Dante Domenella (Stanford University)
It’s a well-documented fact in economics that people in larger cities are more productive. They earn higher wages, produce more patents, and even enjoy better restaurants. This fact even extends to health care, as I show in my paper. In particular, in the context of hip or knee replacements, simply being in a larger market is associated with better health outcomes, as shown in Figure 1. To make this fact concrete, consider the small town of Geneseo, Illinois (population 6,500) versus the Chicago metropolis (population 2.7 million). The probability of a post-surgical complication following a hip or knee replacement is 3.8% in Geneseo but only 2.9% in Chicago.
This pattern is both striking and reflected in other work, but it immediately raises a critical question. Why? Understanding the reason for these differences is crucial for designing public policy, especially place-based policy designed to improve health outcomes in rural areas.
Setting the stage: Do cities attract different people or do they create better outcomes?
Broadly, there are two competing explanations for why people in big cities are more productive: selection or causal mechanisms.
Under the selection explanation, larger markets don't create better outcomes. Instead, they simply attract different types of people. Perhaps healthier, wealthier, or more informed patients are more likely to live in urban areas. Or maybe higher-skilled physicians prefer living in cities because they can make higher profits there. If this latter reason is true, then the policy fix is relatively straightforward: find ways to attract high-quality doctors to smaller towns, such as through loan forgiveness programs or subsidies.
Under the causal mechanism explanation, market size itself creates benefits. That is, when lots of people and firms agglomerate together, productivity increases. For example, in the context of health care, larger markets may facilitate more specialization or knowledge spillovers, which can then improve health outcomes. If these causal mechanisms explain the “big city advantage,” then policy should perhaps leverage the benefits of scale, perhaps by subsidizing patients to travel to these bigger cities.
Exploring specialization as a causal mechanism
In my job market paper, I investigate one specific causal mechanism—specialization—that may generate benefits for patients in urban markets in the context of hip and knee replacements. The idea is best summed up in Adam’s Smith’s famous observation that "the division of labor is limited by the extent of the market." Larger cities facilitate more specialization. In turn, greater specialization allows physicians to focus on a narrower range of procedures. My data shows this phenomenon is not just theoretical. As shown in Figure 2, the average surgeon in the town of Geneseo, Illinois performs 22 hip and knee replacements per year but 43 per year in Chicago.
Of course, this story only holds water if surgeons who perform more procedures actually generate better health outcomes for their patients. A large literature suggests that this fact is indeed true, and Figure 3 corroborates these findings. But, as any economist will tell you, this relationship may not be causal. It could be, for example, that the more skilled surgeons are in high demand and therefore naturally have higher volume.
Therefore, to causally estimate the relationship between surgeon volume and patient outcomes, I use differential distance as an instrument. The idea is that some surgeons have higher volume because they practice closer to patients relative to other surgeons. By comparing otherwise similar surgeons who have different volume due to their relative distance to surgeons, I can isolate the true “practice makes perfect” effect.
I show that, in fact, practice does make perfect, as patients treated by higher-volume surgeons have better health outcomes. Through sheer repetition, high-volume surgeons and their teams become masters of the entire process. This estimate, though, is only about half the size of the raw correlation. This result tells a crucial story—both selection and causal forces are at play. The more skilled surgeons do perform more hip and knee replacements, but the act of performing them makes them even better.
So, how much of the "big city advantage" can be explained by this specialization mechanism? My estimates indicate that it explains 22% of the benefits of market size.
From intellectual curiosity to policy
If a key problem is that patients are often treated by lower-volume surgeons, especially in small markets, what can policy do about it? I use my model to simulate the gold standard policy, an “optimal” policy, and three prominent policies.
The “gold standard”: An optimal policy
I first calculate an optimal policy, which is essentially the best any policy could be. What is the value of this theoretical policy? First, it gives me a benchmark to compare the prominent policies to it. Second, it shows me which surgeons are the high-volume surgeons, so I know how to better target the prominent policies if they are not that successful. This gold standard essentially sets surgeon-specific subsidies and taxes that perfectly guide every patient to the best possible surgeon.
I show that the optimal policy generates large welfare gains that come almost entirely through improved health outcomes. The optimal policy achieves these gains by introducing large subsidies for high-quality surgeons.
Prominent policy 1: Minimum volume standard
Moving to the real world, one prominent policy is a "minimum volume standard," which effectively only allows patients to choose surgeons who meet a certain annual volume threshold. This policy already exists for stroke and cardiovascular procedures in some states.
I show that this policy generates sizable welfare gains. The policy unsurprisingly steers patients toward higher-volume surgeons. Since these surgeons are also innately more skilled, health outcomes improve both because these surgeons are higher quality and because they improve even more with the additional volume. However, this policy only achieves 7% of the gain from the optimal policy, suggesting that more targeted policies may be better. Additionally, it disproportionately harms rural patients, as they are forced to travel farther distances.
Prominent policies 2 and 3: Subsidizing patient transportation and moving surgeons to government-designated shortage areas
What happens when I evaluate other prominent policies, like subsidizing patient transportation costs or moving surgeons to government-designated shortage areas?
Unfortunately, these two other prominent policies are far less effective than the minimum volume standard. Because patients can use the subsidies for patient transportation to travel to any surgeon, they are not targeted enough to generate large welfare effects. Meanwhile, moving surgeons simply does not generate strong enough incentives for patients to substantially change who they choose.
Policy conclusions
The choice of surgeon is heavily shaped by economic geography. In this paper, I show that large markets create better health outcomes, in significant part because they enable the specialization that turns a good surgeon into a great one.
I then leave you with three specific policy implications:
- Targeted concentration can improve outcomes in two ways: Targeting concentration among high-quality surgeons improves outcomes both because these surgeons are innately higher quality and because the increased volume makes them even better.
- Introduce large and targeted changes to patient choice: Unlike the other two policies, the minimum volume standard is targeted at certain surgeons and forces patients to choose new surgeons.
- Target quality: The prominent policies only achieve a small fraction of the optimal policy. As shown in the optimal policy, targeting quality may therefore yield much larger gains.
About the Author
Dante Domenella a PhD candidate in economics at Stanford University.
His research interests are health, public, and labor economics. To learn more about his research, visit: https://sites.google.com/view/ddomenella/
Social Media Handles:
