China Finance Research Conference�Beijing, 2023/7

Collateral Constraints and Asset Prices:

Evidence from Structured Funds

Wei Li, Greg Phelan, and Yongqin Wang

​

Discussant：Zhuo Chen

1

Motivation and Overview

• This paper examines how leverage affects asset prices using the Chinese structured A-B funds.
• Findings:
• Leveraged B fund premium, defined as the difference between the traded B-share price and NAV is positively related to B fund leverage.
• Besides time series and cross-sectional variation in B-share’s effective leverage due to parent fund NAV changes, regular and irregular restructurings are used as exogeneous shocks to fund leverage, which also suggest asset leverage is priced in B fund.
• Economic-wise, a one standard deviation increase in leverage causes the collateral value to increase by 2.14%-3.5%
• The A-share leverage premium (earned by A-share investors by providing leverage to constrained B-share investors) is robust after controlling for potential default risk.
• The findings are consistent with a stylized model similar to Geanakoplos (2003).
• Contribution:
• A new piece of evidence supporting the asset pricing impacts of leverage.
• Fundamental-free variations in leverage lead to clear causal relationship.

2

Comment 1: The source of leverage premium variations

• Is the documented relationship between leverage and B-fund premium time series or cross sectional?
• Time series: for the same fund, increase in the NAV causes B-fund leverage to decrease
• Cross-sectional: for the same day, relative performance of parent fund NAV leads to leverage difference
• By adding different sets of fixed effects, the authors can pin down the source of such variations
• I would suggest to only focus on the passive parent fund, e.g., those tracking HS300 index, to rule out other impacts from active management
• For the set of passive structured funds tracking the same index, their leverage is related to the number of days after any restructuring, which can be used as an instrumental variable.
• Unobserved fund-day shocks may affect the A/B premium for non-passive funds.
• Leverage premium is time-varying and could be related to market-wide conditions.
• A-fund leverage premium is just the other side of the same coin, i.e., the premium earned by A-fund institutional investors (by purchasing A shares at a discount of their NAV)
• Without limits to arbitrage, the parent funds premium should be zero.
• Why do the estimated values of leverage differ when using A fund premium and B fund premium? So does the observed difference only reflect time-varying arbitrage constraints?

​

​

​

​

3

Comment 2: Other forces driving of the B-share premium

• B-fund investors are mostly naïve retail investors who are possibly borrowing constrained.
• But leverage constraint is not the only reason they want to hold B funds
• B-share investors made mistakes, especially before the descending restructuring was about to happen
• They mistakenly think they just invest in another type of index fund due to misleading marketing information of fund management companies
• Alternatively, they could use HS300 futures to lever up, but did they do that?
• Suggestions:
• Drop those descending restructuring as the B share premium reflects not just leverage-induced price appreciation but also investor mistakes.
• Exploit account-level trading data of B-fund investors on their other types of margin trading

​

​

​

​

​

​

4

Comment 3: Additional comments

• Need provide more evidence on the size of structured fund market as a fraction of the underlying indices (especially the HS300 index) to support the argument that the aggregate market valuation cannot be affected by structured funds’ flows.
• Provide the distribution of funds trading different indices.
• Regular restructuring may provide more clear identification for exogeneous variation in leverage.
• Other related literature: Lu and Qin, Leveraged Funds and the Shadow Cost of Leverage Constraints, The Journal of Finance, 2021

​

5