Sergei Glebkin, Ji Shen, Bart Yueshen
Based on: Review of Financial Studies, 36(2): 571-614 DOI: https://doi.org/10.1093/rfs/hhac009
Recent advances in technology have led to a profound transformation of asset markets. Among those affected is the multi-trillion-dollar US corporate bond market.
Traditionally, it was structured as a classical over-the-counter (OTC) market, where a customer willing to buy a particular bond searches for a dealer willing to sell it by picking up the phone and calling different dealers. Once a dealer is found, the price is determined via bilateral negotiation between the customer and the dealer.
But now, modern technology allows customers to contact dealers electronically. Moreover, they can contact multiple dealers at the same time. Instead of negotiating the price, the customer can rely on the competition among contacted dealers and select the most attractive offer. This new way of matching customers and dealers is called simultaneous multilateral search (SMS).
Currently, the old (“voice”) market structure coexists with SMS in corporate bonds and many other OTC markets. Will the voice market become extinct? Is the SMS market structure better? What are the implications of introducing SMS?
At first glance, SMS seems to be a better market structure: It allows customers to contact more dealers and, by fostering competition among dealers, allows customers to get a better deal. In practice, the adoption of SMS has been quite sluggish: The old technology is still used alongside the new one, especially in times of distress.
Should policymakers encourage the transition to the new market structure? Not necessarily. The goal is to achieve the optimal allocation of resources. In our case, this means transferring bonds from those who value them the least, i.e., customer-sellers, to those who value them the most, i.e., customer-buyers. The key difference between voice technology and SMS is search capacity—the number of dealers a customer can contact simultaneously. Voice has a smaller search capacity than SMS.
SMS might not always outperform voice in allocation efficiency because changes in search capacity have an asymmetric impact on different sides of the market. Consider a market for a specific corporate bond that is in excess supply: Most dealers have the bond in their inventory, while only a few do not. Consequently, when a customer wants to buy the bond, it’s easy to find a dealer willing to sell since most dealers have it available. But when a customer wants to sell the bond, they may struggle to find a dealer interested in buying it, as fewer dealers need it.
Transitioning from voice to SMS increases search capacity, but this only marginally improves the already high chances of customer-buyers finding matching dealers. In contrast, it can significantly enhance the odds that sellers will find a match by allowing them to reach out to more dealers. This asymmetric increase in matching probabilities—significant for sellers and marginal for buyers—creates an imbalance: The increased inflow of bonds into the dealer sector isn’t matched by a comparable outflow. As a result, bonds accumulate with dealers, leading to a bottleneck and an inefficiency because these assets would be more beneficial if transferred to customer-buyers.
We also examine the effects of enhanced transparency, where disclosing past trades (through trade-reporting systems like TRACE in the corporate bond market) gives customers a clearer understanding of which dealers hold specific assets. This allows them to direct their searches more efficiently. We demonstrate that a similar bottleneck can arise when transparency improves: As customers direct their searches more accurately, the matching on the short and on the long side improves asymmetrically, hindering the efficient passing of the asset through dealers.
The overall effect in the bottleneck scenarios described above is that the asset becomes “clogged” with the dealers rather than flowing efficiently from sellers to buyers. This creates a trade-off in terms of overall welfare. The reduction in the number of unmatched sellers has a positive effect, but it is offset by the negative impact of increasing the number of unmatched buyers. In slow markets where assets change hands infrequently, the bottleneck becomes significant, and the overall impact on welfare is negative. Conversely, in fast markets, the bottleneck is minimal (as trade facilitates the flow of assets), resulting in a positive overall welfare effect. Such a distinction between fast- and slow-moving markets is important. While most liquid corporate bonds on SMS can trade within minutes, clearing the SMS protocols for collateralized loan obligations (CLOs) may take a day or more. The impact of transitioning from voice to SMS and that of increasing transparency is not universal and varies depending on the type of asset.
Recent advances in technology have led to a profound transformation of asset markets. Among those affected is the multi-trillion-dollar US corporate bond market.
Traditionally, it was structured as a classical over-the-counter (OTC) market, where a customer willing to buy a particular bond searches for a dealer willing to sell it by picking up the phone and calling different dealers. Once a dealer is found, the price is determined via bilateral negotiation between the customer and the dealer.
But now, modern technology allows customers to contact dealers electronically. Moreover, they can contact multiple dealers at the same time. Instead of negotiating the price, the customer can rely on the competition among contacted dealers and select the most attractive offer. This new way of matching customers and dealers is called simultaneous multilateral search (SMS).
Currently, the old (“voice”) market structure coexists with SMS in corporate bonds and many other OTC markets. Will the voice market become extinct? Is the SMS market structure better? What are the implications of introducing SMS?
At first glance, SMS seems to be a better market structure: It allows customers to contact more dealers and, by fostering competition among dealers, allows customers to get a better deal. In practice, the adoption of SMS has been quite sluggish: The old technology is still used alongside the new one, especially in times of distress.
Should policymakers encourage the transition to the new market structure? Not necessarily. The goal is to achieve the optimal allocation of resources. In our case, this means transferring bonds from those who value them the least, i.e., customer-sellers, to those who value them the most, i.e., customer-buyers. The key difference between voice technology and SMS is search capacity—the number of dealers a customer can contact simultaneously. Voice has a smaller search capacity than SMS.
SMS might not always outperform voice in allocation efficiency because changes in search capacity have an asymmetric impact on different sides of the market. Consider a market for a specific corporate bond that is in excess supply: Most dealers have the bond in their inventory, while only a few do not. Consequently, when a customer wants to buy the bond, it’s easy to find a dealer willing to sell since most dealers have it available. But when a customer wants to sell the bond, they may struggle to find a dealer interested in buying it, as fewer dealers need it.
Transitioning from voice to SMS increases search capacity, but this only marginally improves the already high chances of customer-buyers finding matching dealers. In contrast, it can significantly enhance the odds that sellers will find a match by allowing them to reach out to more dealers. This asymmetric increase in matching probabilities—significant for sellers and marginal for buyers—creates an imbalance: The increased inflow of bonds into the dealer sector isn’t matched by a comparable outflow. As a result, bonds accumulate with dealers, leading to a bottleneck and an inefficiency because these assets would be more beneficial if transferred to customer-buyers.
We also examine the effects of enhanced transparency, where disclosing past trades (through trade-reporting systems like TRACE in the corporate bond market) gives customers a clearer understanding of which dealers hold specific assets. This allows them to direct their searches more efficiently. We demonstrate that a similar bottleneck can arise when transparency improves: As customers direct their searches more accurately, the matching on the short and on the long side improves asymmetrically, hindering the efficient passing of the asset through dealers.
The overall effect in the bottleneck scenarios described above is that the asset becomes “clogged” with the dealers rather than flowing efficiently from sellers to buyers. This creates a trade-off in terms of overall welfare. The reduction in the number of unmatched sellers has a positive effect, but it is offset by the negative impact of increasing the number of unmatched buyers. In slow markets where assets change hands infrequently, the bottleneck becomes significant, and the overall impact on welfare is negative. Conversely, in fast markets, the bottleneck is minimal (as trade facilitates the flow of assets), resulting in a positive overall welfare effect. Such a distinction between fast- and slow-moving markets is important. While most liquid corporate bonds on SMS can trade within minutes, clearing the SMS protocols for collateralized loan obligations (CLOs) may take a day or more. The impact of transitioning from voice to SMS and that of increasing transparency is not universal and varies depending on the type of asset.
Assistant Professor of Finance
Lee Kong Chian School of Business
Singapore Management University
