Do Intermediaries Matter for Aggregate Asset Prices? by Valentin Haddad and Tyler Muir Discussion Pietro Veronesi The University of Chicago Booth School of Business
Main Contribution and Outline of Discussion Main contribution of the paper: 1. Provide a (simple) testable prediction of the relative impact of intermediaries on the risk premium of intermediated asset classes More intermediation = Higher elasticity of risk premium to changes in risk aversion 2. Provide (some) evidence supporting the prediction
Main Contribution and Outline of Discussion Main contribution of the paper: 1. Provide a (simple) testable prediction of the relative impact of intermediaries on the risk premium of intermediated asset classes More intermediation = Higher elasticity of risk premium to changes in risk aversion 2. Provide (some) evidence supporting the prediction Outline of discussion 1. Frictionless Model and Institutional Leverage 2. Model and Interpretation 3. Questions on Empirics 4. Final comments
Frictionless Dynamic Economy Frictionless model with many agents each born with a tree dd it D it = µ i,t dt + σ i,t db t + σ i,i,t db it
Frictionless Dynamic Economy Frictionless model with many agents each born with a tree dd it D it = µ i,t dt + σ i,t db t + σ i,i,t db it Each tree has a idiosyncratic shock db it.
Frictionless Dynamic Economy Frictionless model with many agents each born with a tree dd it D it = µ i,t dt + σ i,t db t + σ i,i,t db it Each tree has a idiosyncratic shock db it. Agents don t like idio risk = set up firm, issue equity P i (t) (claim to D i,t )
Frictionless Dynamic Economy Frictionless model with many agents each born with a tree dd it D it = µ i,t dt + σ i,t db t + σ i,i,t db it Each tree has a idiosyncratic shock db it. Agents don t like idio risk = set up firm, issue equity P i (t) (claim to D i,t ) Agents maximize intertemporal utility [ ] max E 0 e ρt u i (C it, X it, t) C it subject to budget constraint. Initial wealth equal value of the tree at time 0. Invest in aggregate stock Borrow or lend depending on risk aversion 0
Frictionless Dynamic Economy 2 Intermediaries are passive: Just facilitate risk transfers. Assets: 1. Underwrite (purchase) equity P i (t) from individual firms 2. Make loans L i (t) to households to help finance their dream consumption
Frictionless Dynamic Economy 2 Intermediaries are passive: Just facilitate risk transfers. Assets: 1. Underwrite (purchase) equity P i (t) from individual firms 2. Make loans L i (t) to households to help finance their dream consumption Liabilities 1. Issue equity S j (t) to market 2. Receive deposits from households D i (t)
ó ^ Intermediation in a Frictionless Economy ª «± ² ³ µ ~ ƒ ˆ Š Œ Ž ¹ º» ¼ ½ ¾ À Á  ½ ¾ š œ ž Ÿ. / 0 1 2 3 4 5 6 7 8 9 : ; < = >? @ A B C C C C C D E F G H I J K L M N O P Q R Ã Ä Å Æ Ç È É Ê Ê Ë Ì Í Î Ï Ð Ñ Ò Ó Ô Õ Ô Ö Ø Ù Ú Û Ü Ý Þ ß à á â ã ä å æ ç è é ê ë ì í î ï ð ñ ò ô õ ö ø ù ú ú û ü ý þ ÿ ÿ ÿ! " # $ % & ' ( ) * +, - R S T T U V W X Y Z Z [ \ ] _ ` ` a b c d e f f g h i j k l l l l l m n o p q q r s t u v w x y z z { } & ' ( ' ) * +, -.. / 0 1 2 3 4 3 5 6 7 8 9 : 9 ; < = >? @ @ @ @ @ A B C D E D F G H I J K L M N M O P Q! " # $ %
Differential Variation in Risk Aversion Suppose agents risk aversion differentially changes over the business cycle = Motives for trade = Change allocation to intermediary equity, deposits, and loans
Differential Variation in Risk Aversion Suppose agents risk aversion differentially changes over the business cycle = Motives for trade = Change allocation to intermediary equity, deposits, and loans In a heterogeneous habit model, Santos and Veronesi (2017) show: risk tolerant agents risk averse agents Intermediary good times stock holdings stock holding equity value debt deposits (overnight) debt debt/mkt equity bad times stock holdings stock holding equity value debt deposits (overnight) debt debt/mkt equity
Intermediaries Debt and Debt/Equity in a Simulation Run 0.06 A. Surplus Consumption Ratio 6 B. Economic Uncertainty 0.04 0.02 Percent 4 2 0 0 50 100 150 200 250 300 350 400 Calendar Time (Quarters) C. Price / Dividend Ratio 40 0 0 50 100 150 200 250 300 350 400 Calendar Time (Quarters) D. Return Volatility 60 20 Percent 40 20 0 0 50 100 150 200 250 300 350 400 Calendar Time (Quarters) 0.95 E. Aggregate Debt/Output and Stock Holdings 0.44 0 0 50 100 150 200 250 300 350 400 Calendar Time (Quarters) F. Aggregate Debt/Wealth 0.4 0.9 0.43 0.2 Debt/Output Stocks 0.85 0.42 0 50 100 150 200 250 300 350 400 Calendar Time (Quarters) 0 0 50 100 150 200 250 300 350 400 Calendar Time (Quarters)
Intermediary Leverage is just a Proxy of Aggregate Risk Aversion In the model l t = Q (Aggregate Risk Aversion) Therefore: RP i = Cov (R i, d log(c)) }{{} log-utility RP +λ(l t ) Cov (R i, dl t ) }{{} Intermediary Leverage RP
Intermediary Leverage is just a Proxy of Aggregate Risk Aversion In the model l t = Q (Aggregate Risk Aversion) Therefore: RP i = Cov (R i, d log(c)) }{{} log-utility RP +λ(l t ) Cov (R i, dl t ) }{{} Intermediary Leverage RP Intermediary s leverage is priced because it proxies for aggregate risk aversion If other proxies of aggregate risk aversion are noisier, then intermediary leverage may work better
Intermediary Leverage is just a Proxy of Aggregate Risk Aversion In the model l t = Q (Aggregate Risk Aversion) Therefore: RP i = Cov (R i, d log(c)) }{{} log-utility RP +λ(l t ) Cov (R i, dl t ) }{{} Intermediary Leverage RP Intermediary s leverage is priced because it proxies for aggregate risk aversion If other proxies of aggregate risk aversion are noisier, then intermediary leverage may work better Model produces several stylized facts: Intermediaries 1. debt in good times 2. debt/mkt equity in bad times 3. Market leverage has negative risk price 4. Book leverage has positive risk price
Intermediary Asset Pricing: Book and Market Leverage Table 3: The Market Price of Leverage Risk Panel A - Data (FF25) α 3.19 0.76 1.07 (3.05) (0.62) (0.97) Market Return -0.89 0.97 0.82 (-0.72) (0.69) (0.61) Market Leverage -0.22 (-2.13) Book Leverage 0.63 (3.07) R2 (%) 6.54 50.77 53.35 Panel B - Model (Individual Trees) α 0.02 0.10 0.08 Market Return 2.05 1.96 1.98 Market Leverage -0.04 Book Leverage 0.03 Market leverage and book leverage have different risk prices in data and model
Intermediary Asset Pricing: Book and Market Leverage Table 4: The Predictability of Aggregate Stock Returns Panel A. Predictability with Book Leverage. Data 1 year 2 year 3 year 4 year 5 year Coef ( 100) -1.78-1.79-2.17-3.13-9.89 (-0.83) (-0.72) (-0.89) (-1.03) (-3.29) R2 0.01 0.01 0.01 0.01 0.07 Panel B. Predictability with Market Leverage. Data Coef ( 100) 3.66 6.21 8.56 10.03 13.06 (1.57) (1.50) (2.18) (2.51) (3.84) R2 0.04 0.07 0.10 0.12 0.19 Panel C. Predictability with Book Leverage. Model Coef ( 100) -3.57-7.28-10.49-12.62-14.13 (-3.45) (-3.09) (-3.18) (-3.34) (-3.52) R2 0.02 0.05 0.08 0.10 0.12 Panel D. Predictability with Market Leverage. Model Coef ( 100) 5.86 10.91 14.69 17.35 19.36 (8.08) (7.69) (7.55) (7.54) (7.74) R2 0.06 0.12 0.17 0.20 0.22 Market leverage and book leverage predict future returns with different signs in data and model
Haddad and Muir Idea Intermediaries balance sheet expand and decline even without any friction
Haddad and Muir Idea Intermediaries balance sheet expand and decline even without any friction But there are frictions = intermediaries may have additional impact on prices
Haddad and Muir Idea Intermediaries balance sheet expand and decline even without any friction But there are frictions = intermediaries may have additional impact on prices Impact should be especially visible in assets that are more intermediated
Haddad and Muir Idea Intermediaries balance sheet expand and decline even without any friction But there are frictions = intermediaries may have additional impact on prices Impact should be especially visible in assets that are more intermediated Idea: Rank assets in order of household cost of trading them A change in intermediary risk aversion = larger impact on risk premium of more intermediated assets In their model RP i /RP i log(intermediary RA) household trading cost
Implementation: Rankings and Risk Aversion Use data to rank asset classes in order of intermediation Table 9: Ranking of Asset Classes. Ranking by degree of intermediation by source, with our chosen ranking on the top row. From left to right is less intermediated asset classes, with relatively easier access to investing by households, to more intermediated asset classes, with lower participation by households. The sources for the rankings are: the Flow of Funds (FoF), BIS derivatives positions, Vale-at-Risk (VaR), and ETF expense ratios. The text explains these sources and rankings in detail. Our Ranking Stocks Bonds Options Sov Bonds Comm FX CDS FoF Stocks Bonds Sov Bonds VaR Stocks Bonds Comm FX BIS Bonds Options Comm FX CDS Expense Stocks Bonds Sov Bonds FX Comm Options CDS
Implementation: Rankings and Risk Aversion Use data to rank asset classes in order of intermediation Table 9: Ranking of Asset Classes. Ranking by degree of intermediation by source, with our chosen ranking on the top row. From left to right is less intermediated asset classes, with relatively easier access to investing by households, to more intermediated asset classes, with lower participation by households. The sources for the rankings are: the Flow of Funds (FoF), BIS derivatives positions, Vale-at-Risk (VaR), and ETF expense ratios. The text explains these sources and rankings in detail. Our Ranking Stocks Bonds Options Sov Bonds Comm FX CDS FoF Stocks Bonds Sov Bonds VaR Stocks Bonds Comm FX BIS Bonds Options Comm FX CDS Expense Stocks Bonds Sov Bonds FX Comm Options CDS Find proxies of intermediary Risk Aversion Book leverage (AEM) and market leverage (HKM) GZ spread Find proxies of households Risk Aversion cay
Main Result Table 1: Main predictive regressions. Predictive regressions of future excess returns in each asset class on our proxy for intermediary risk aversion, γ Int. Our proxy is the average of the standardized versions of the AEM and HKM intermediary factors. We run: r i,t+1 /E[r i,t+1 ] = a i + b i x t + ε i,t+1 and report b i which gives the elasticity of the risk premium of asset i to x. See text for more details. Bootstrapped standard errors are in parenthesis and adjust for the fact that unconditional expected returns (E[r i,t+1 ]) are estimated. See text for more details. Panel A: Quarterly Returns (1) (2) (3) (4) (5) (6) (7) Stocks Bonds Sovereign Commodities CDS Options FX γ Int -0.75-0.49-1.06-3.95-2.68-1.43-0.45 (0.50) (0.26) (0.43) (1.51) (0.77) (0.61) (0.22) N 167 148 65 105 47 103 116 R 2 1.6% 1.5% 15.4% 5.4% 35.6% 4.4% 3.2% Panel B: Annual Returns (1) (2) (3) (4) (5) (6) (7) Stocks Bonds Sovereign Commodities CDS Options FX γ Int -0.34-0.35-0.64-2.52-1.08-0.68-0.22 (0.27) (0.15) (0.16) (0.78) (0.44) (0.30) (0.09) N 164 145 62 102 44 100 113 R 2 1.5% 2.7% 26.2% 7.1% 23.1% 3.8% 3.4%
Main Result Table 1: Main predictive regressions. Predictive regressions of future excess returns in each asset class on our proxy for intermediary risk aversion, γ Int. Our proxy is the average of the standardized versions of the AEM and HKM intermediary factors. We run: ri,t+1 /E[ri,t+1 ] = ai + bi xt + εi,t+1 and report bi which gives the elasticity of the risk premium of asset i to x. See text for more details. Bootstrapped standard errors are in parenthesis and adjust for the fact that unconditional expected returns (E[ri,t+1 ]) are estimated. See text for more details. Panel A: Quarterly Returns (1) (2) (3) (4) (5) (6) (7) Stocks Bonds Sovereign Commodities CDS Options FX Ranking (1) (2) (4) (5) (7) (3) (6) γ Int -0.75 (0.50) -0.49 (0.26) -1.06 (0.43) -2.68 (0.77) -1.43 (0.61) -0.45 (0.22) N R2 167 1.6% 148 1.5% 65 105 47 15.4% 5.4% 35.6% Panel B: Annual Returns (3) (4) (5) Sovereign Commodities CDS 103 4.4% 116 3.2% (1) Stocks (2) Bonds (6) Options (7) FX γ Int -0.34 (0.27) -0.35 (0.15) -0.64 (0.16) -2.52 (0.78) -1.08 (0.44) -0.68 (0.30) -0.22 (0.09) N R2 164 1.5% 145 2.7% 62 26.2% 102 7.1% 44 23.1% 100 3.8% 113 3.4% -3.95 (1.51) from Table 9
Ranking? Ranking captures HH s cost of trading But it could be correlated with other characteristics Implicit Leverage, Beta, Illiquidity, Other (true) risk factors Intermediaries may choose to trade among each other specific assets. E.g., they may trade CDS to share risk among each other from loans
Ranking? Ranking captures HH s cost of trading But it could be correlated with other characteristics Implicit Leverage, Beta, Illiquidity, Other (true) risk factors Intermediaries may choose to trade among each other specific assets. E.g., they may trade CDS to share risk among each other from loans In simulations: Elasticity of Risk Premium vs Implicit Leverage 6.5 6.4 6.3 Elasticity 6.2 6.1 6 5.9 5.8-0.8-0.6-0.4-0.2 0 0.2 0.4 0.6 0.8 Implicit Leverage of Asset
Intermediaries Risk Aversion vs AEM, HKM, GZ, etc. Haddad and Muir model has specific predictions for the elasticity of risk premium on intermediaries risk aversion
Intermediaries Risk Aversion vs AEM, HKM, GZ, etc. Haddad and Muir model has specific predictions for the elasticity of risk premium on intermediaries risk aversion AEM and HKM hardly proxy for institutional risk aversion, as they are related to book or market leverage. What is the prediction about risk premium and intermediaries leverage? What is the prediction about risk premium and cay (consumption decisions)?
Final Comment The question is interesting and important Not swayed yet by the evidence. Need more.