Healthcare price transparency has been a focal point in U.S. policy debates due to large price variations within procedure amid escalating healthcare costs. Many contend that insufficient transparency contributes to the nation’s high healthcare expenditures. The complexity of price transparency stems from the fact that key healthcare prices, known as the “allowed amount”, are determined through private negotiations between insurers and providers, who are often restricted from disclosing these negotiated rates. Early efforts to enhance transparency began with a requirement for providers to display their chargemaster data online, which was soon recognized as being substantially different from the actual prices faced by patients. This realization spurred additional initiatives at the employer, insurer, and state levels to mandate the disclosure of negotiated prices and to develop tools that allow beneficiaries to search for these prices. The most significant recent development is the 2021 Hospital Price Transparency Final Rule, which mandates a customer-friendly display of allowed amount information for a select list of popular elective procedures.

Despite these efforts, the question remains: Can price transparency effectively reduce healthcare costs? There are concerns that price transparency might not lead to lower healthcare prices due to two main factors. First, healthcare consumers often show considerable price inelasticity, suggesting that even with price information, patient behavior might not change enough to impact costs (Desai et al., 2021). Second, widespread transparency might inadvertently facilitate coordination and collusion among providers, leading to higher, not lower, prices (Cutler and Dafny, 2011). Prior research finds mixed empirical effects: some studies show overall price reductions (Whaley et al., 2014, Wu et al., 2014, Sinaiko et al., 2016, Desai et al., 2017, Lieber, 2017, Brown, 2019a, Brown, 2019b, Whaley, 2019, Knies, 2021), while others find no effect or even price increases (Desai et al., 2016, Desai et al., 2017, Lieber, 2017, Allcott et al., 2021, Christensen et al., 2020).1 This paper leverages a recent statewide price disclosure to provide new evidence on how price transparency affects healthcare prices. It examines both the role of provider–insurer negotiations and patient responses in driving these effects.

To identify the effect of price transparency, I use a natural experiment of healthcare price disclosure on a variety of medical procedures for providers and insurers in New Hampshire. Prices of a group of popular, elective procedures were posted on a state run website using staggered timing over years, while prices of other medical procedures remained undisclosed. I employ difference-in-differences estimators, leveraging variations in the timing of price disclosures and the selection of procedures made public. The price posted is the allowed amount between providers and insurers. The website’s structure closely aligns with the prevalent features of current price transparency initiatives—it offers the provider and insurer specific allowed amounts for some popular elective procedures. Therefore, the findings of this study can provide valuable insights for other price transparency initiatives.

This study relies on two primary data sources. The first is the price transparency website data, which includes URL-specific website traffic, the displayed prices for each procedure by insurer and provider, and the timing of each price listing. The second is New Hampshire’s all-payer claims data, including all medical claims from commercially insured patients within the state. The analysis focuses on two categories of medical procedures newly listed on the website: surgical and radiology procedures.2

I first assess the website’s effectiveness by comparing website displayed providers and insurers with those in claims data, evaluating its initial impact on price-seeking patients. I find that for each procedure, the website encompasses price information for approximately 20% of the providers and half of the insurers offering that specific procedure. In particular, the website mainly includes high-volume providers and insurers but leaves out a substantial portion of lower-volume ones. While prior price transparency literature found limited engagement with price tools, my assessment shows limited coverage by the website, which may help explain the low rates of patient use reported in prior studies. (Whaley et al., 2014, Wu et al., 2014, Desai et al., 2016, Sinaiko et al., 2016, Lieber, 2017, Desai et al., 2017, Brown, 2019b, Whaley, 2019, Knies, 2021)

A healthcare price transparency tool can influence prices through two channels. From the perspective of providers and insurers, it could affect the price-setting process of healthcare procedures.3 For instance, if price transparency promotes competition among providers, they might set lower prices during negotiations. Conversely, if a price transparency tool fosters collusion among providers, they might set higher prices. I refer to this channel as the “bargaining effect” of price transparency. From the patient’s perspective, the price transparency tool provides price information to incentivize them to seek care from different providers. I refer to this channel as the “patient sorting effect” of price transparency. Each channel offers different empirical predictions that I use to inform the empirical analysis of the mechanisms.

My paper primarily examines the bargaining effects of a price transparency tool. A smaller subset of existing studies has examined the impact of price transparency on providers and insurers, often due to data limitations. Among these studies, findings range from no impact to a negative impact on the negotiated amounts across various procedures and settings (Wu et al., 2014, Brown, 2019a, Brown, 2019b, Whaley, 2019). My setting is particularly well-suited for such an analysis for two reasons. First, the price transparency initiative under study is regionally substantial, making it likely to elicit meaningful responses from insurers and providers. Second, access to the state all-payer claims data enables observation of all negotiated allowed amounts across providers, insurers, and procedures, allowing me to analyze within provider–insurer–procedure variations in allowed amounts before and after price disclosure.

I find that providers and insurers significantly reduced allowed amounts in response to price disclosure for certain procedures. Specifically, negotiated allowed amounts decreased by 5.1% for surgical procedures and 9.1% for radiology procedures, while the evidence is mixed and does not point to a clear effect for lab procedures.4 These results underscore the heterogeneous impact of price transparency across different categories of medical procedures. On the one hand, surgical procedures, traditionally considered relatively non-standard, still experienced a significant decrease in allowed amounts following price disclosure. On the other hand, lab procedures, which have the lowest average allowed amount and website utilization rate among the categories, showed no clear response to price transparency. An event study model further reveals no significant differences in the pre-disclosure price trends between procedures that were eventually added to the website and those that were not. This finding supports a causal interpretation of the website’s impact on negotiated prices.

Next, I examine the spill-over effect of the website. Despite the website not displaying prices for all insurers and providers for each procedure, the effect of price transparency on allowed amounts remains robust across all providers and insurers, whether or not their prices were posted. Additionally, I observe significant reductions in allowed amounts for providers in adjacent states, who are direct competitors of New Hampshire providers, following the price disclosure. This suggests that by disclosing prices of only high-volume providers and insurers, the allowed amounts for all providers and insurers, including those in adjacent states, significantly decreased.

In addition to analyzing the bargaining effect, the paper investigates the mechanisms behind price transparency by examining the patient-sorting effect. Specifically, I analyze whether patients shift their visits from high-cost to low-cost providers following the disclosure of price information. Results show no evidence of patients switching to lower-cost providers after transparency implementation. Combined with the bargaining effect findings, my paper contributes new evidence showing that the reduction in allowed amount of a healthcare visit following price transparency is primarily driven by providers and insurers negotiating lower amounts.

Finally, I analyze price dispersion. While theory suggests transparency reduces dispersion by lowering search costs (Salop and Stiglitz, 1977, Stigler, 1961), and empirical studies support this in other markets (Sorensen, 2000, Aker, 2010, Chandra and Tappata, 2011), I find no significant change in price dispersion. Both high- and low-cost providers lowered prices similarly.

Overall, this paper finds that price transparency significantly reduced allowed amounts for surgery and radiology procedures, but not for lab procedures. The reductions appear to be primarily driven by provider–insurer negotiations rather than by patient responses. My results provide timely evidence on the effectiveness of a statewide price transparency website, showing its potential to reduce healthcare spending. These findings are highly relevant for policymakers seeking actionable strategies to improve transparency and address rising healthcare costs.

Several prior studies have examined the impact of NH HealthCost. Tu and Lauer (2009) found no significant price changes but reported anecdotal evidence of insurer–provider negotiations. Mehrotra et al. (2014) documented low website usage (1% of state residents). Desai et al. (2021) found that an online ad campaign increased website traffic but did not change patient price shopping. The only paper that studied the impact of price transparency on healthcare prices using the website’s price disclosure as a natural experiment is Brown (2019b), which is closely related to this study. Brown (2019b) found a 3% price reduction for imaging procedures following the website’s 2007 launch.

This paper extends previous research by examining a broader set of procedures on NH HealthCost beyond radiology, including less standardized surgical procedures. While Brown (2019b) analyzed the website’s 2007 launch, this study evaluates the impact of adding new procedures between 2015 and 2017, leveraging a more recent natural experiment. The staggered rollout improves identification, and detailed website data allow for an assessment of price transparency’s effects on insurers and providers whose prices were not posted, providing further insights into the website’s effectiveness.

This study also relates to price transparency research in industries other than medical procedures, such as supermarkets (Ater and Rigbi, 2023), gasoline (Rossi and Chintagunta, 2016, Luco, 2019), and medical devices (Grennan and Swanson, 2020). Unlike these markets, healthcare involves quality variation and limited consumer price exposure due to insurance. Nonetheless, my findings show transparency can significantly affect prices despite these complexities.

The paper is organized as follows. Section 2 introduces the basics of the NH HealthCost website and its background. Section 3 outlines mechanisms through which a price transparency website can affect healthcare prices. Section 4 discusses the data used in this study. Section 5 describes the difference-in-differences method employed by this paper. Sections 6 First stage - Provider and insurer availability on NH Healthcost, 7 Results details and discusses the results, while Section 8 concludes.