Gábor Pintér

Are larger trades more or less expensive to execute in bond markets than smaller trades? This is an old and unsettled question in the literature on financial markets. The aim of this blog post is to provide novel answers to this question, based on our recent research using transaction-level data from the UK government and corporate bond markets, over the period 2011–17.[1]

What does previous research say about the size-cost relation?

The existing empirical evidence shows that larger trades incur lower trading costs (‘size discount’) in various over-the-counter (OTC) financial markets such as the market for government bonds and corporate bonds among many others. The size discount is consistent with theories of bilateral trading with imperfect competition. They predict that larger trades get more favourable prices because dealers’ bargaining power decreases in the size of their clients, and larger clients tend to trade larger amounts. However, theories of information asymmetry and inventory imbalances predict ‘size penalty’, in that larger trades would be executed at less favourable prices. That is because of dealers’ fear of being adversely selected by the informed clients or because of dealers’ additional inventory costs when managing a larger trade. To sum up, theories of bargaining give contrasting predictions on the size-cost relation compared to theories of informational asymmetry and inventory imbalances. This has generated some tension in the literature.

New evidence on the size-cost relation

Our research reconciles some of this tension on this literature by providing a new decomposition of the size-cost relation. Specifically, we estimate i) how trading costs vary across clients (‘cross-client variation’) and ii) how trading costs vary across trades of different sizes for the same client (‘within-client variation’). We are able to do this decomposition because our unique dataset contains the identities of counterparties for almost all secondary market transactions in the UK bond markets. This allows us to distinguish between client-specific characteristics (such as traders’ size and type) and transaction-specific characteristics (such as trade size) in determining trading costs.

We use simple panel data techniques, whereby we regress clients’ trading costs on trade sizes and various control variables, using all available client-dealer trades. We compute clients’ trading costs (measured in basis points) as the difference between the transaction prices and a benchmark price (measured as the average transaction price). We find that larger trades get lower trading costs than smaller trades, thereby corroborating the previous literature on the ‘size discount’. However, we find that trading costs increase in trade size once we control for clients’ identities, generating a ‘size penalty’. These two findings are illustrated in Chart 1, which shows the relationship between trade size and trading costs in government bonds from two different model specifications. 

Chart 1: The relation between trade size and trading costs in the gilt market (2011–17)

The left panel of Chart 1 plots the fitted linear regression line from a pooled regression of trading costs on trade size. The trade-level regression shows that larger trades incur lower trading costs, consistent with the findings of size discount in other OTC markets. Our novel contribution is to isolate the within-client variation in the size-cost relation. The right panel of Chart 1 shows the regression line after removing the client-specific average from trading costs and trade size, giving rise to a size penalty. This suggests that the size discount is driven by the cross-client variation, as larger clients with more pricing power are able to gain lower trading costs. The size penalty is driven by the within-client variation of the same trader facing higher trading costs when managing larger trades.

What drives the size penalty?

Our research also provides analysis of the determinants of the size penalty in further detail. For example, we show that the size penalty is larger for hedge funds and asset managers, and it is smaller for pension funds, foreign central banks and insurance companies. Moreover, the size penalty, faced by hedge funds and asset managers, is larger during informationally intensive periods such as trading days that coincide with the arrival of large macroeconomic shocks. In contrast, the size penalty faced by other clients is similar across trading days irrespective of the magnitude of macroeconomic shocks at the time. In addition, the size penalty is larger in corporate bonds than in government bonds, and, importantly, this difference is more pronounced among hedge funds and asset managers compared to other clients.

Our research also compares the size penalty across the UK gilt and US Treasury markets. Given that the US Treasury market is larger, deeper and more liquid than the UK gilt market, both inventory-based theories and information-based theories would predict a smaller size penalty in US Treasuries. The results are consistent with these predictions: the size penalty in US Treasuries is estimated to be about three times as small as in UK gilts.

Taken together, these results are interpreted as evidence that information-based explanations contribute to the heterogeneity in size penalty. To the extent that hedge funds and asset managers are more likely to trade on information than other clients, the differential degree of size penalty across client types, is consistent with theories of asymmetric information. These theories predict that larger trades (especially those of hedge funds) are more likely to be informational, so that dealers rationally charge higher execution costs on these trades in order to be compensated for adverse selection risk.

Why does the size-cost relation matter?

There are various dimensions of market liquidity such as the bargaining power and balance sheet constraints of dealers, the price impact of informed trading among others. Observing the identities of clients (and their transaction prices) in bond markets can help disentangle these different dimensions of liquidity and give us a better understanding on how prices are determined in OTC markets in general.

For example, if one aims to learn about the role of dealers’ market power in shaping market liquidity, then one could compare the transaction prices of different types of clients (eg small retail clients versus large asset managers). If, on the other hand, one aims to understand the price impact of informed trading, then one should compare the transaction prices for small and larger trades initiated by the same smart investor (eg a star hedge fund). Our results show that both channels seem to contribute to variation in bond market spreads.

Gábor Pintér works in the Centre For Central Banking Studies within the Bank.

If you want to get in touch, please email us at [email protected] or leave a comment below.

Comments will only appear once approved by a moderator, and are only published where a full name is supplied. Bank Underground is a blog for Bank of England staff to share views that challenge – or support – prevailing policy orthodoxies. The views expressed here are those of the authors, and are not necessarily those of the Bank of England, or its policy committees.

[1] The results may therefore not be representative of current market liquidity and dynamics.


Source link

You May Also Like

Ally Financial eyes AI-based inventory recommendations | Bank Automation News

Ally Financial is continuing to invest in technology and generative AI to offer…

Zilo nabs $31M in series A | Bank Automation News

Venture funding fell in 2023 as investors took uncertain macroeconomic conditions into…

How to Start a Roth IRA Account in 5 Steps

Chime® is a financial technology company, not a bank. Banking services are…

Standard Chartered: Trump is 'transactional' but may not rock the boat on China

Standard Chartered CEO Bill Winters discusses the outlook for the Chinese economy…