The Value of Control and the Costs of Illiquidity

• Author: ENRIQUE SCHROTH,RUI ALBUQUERQUE
• Date: 01 August 2015
• DOI: http://doi.org/10.1111/jofi.12207
• Published date: 01 August 2015

THE JOURNAL OF FINANCE •VOL. LXX, NO. 4 •AUGUST 2015

The Value of Control and the Costs of Illiquidity

RUI ALBUQUERQUE and ENRIQUE SCHROTH∗

ABSTRACT

We develop a search model of block trades that values the illiquidity of controlling

stakes. The model considers several dimensions of illiquidity. First, following a liq-

uidity shock, the controlling blockholder is forced to sell, possibly to a less efficient

acquirer. Second, this sale may occur at a fire sale price. Third, absent a liquidity

shock, a trade occurs only if a potential buyer arrives. Using a structural estima-

tion approach and U.S. data on trades of controlling blocks of public corporations,

we estimate the value of control, blockholders’ marketability discount, and dispersed

shareholders’ illiquidity-spillover discount.

HIGH OWNERSHIP CONCENTRATION is a predominant phenomenon in the corpo-

rate world. In many countries, including the United States, evidence suggests

that high ownership concentration is pervasive in public corporations.1By

definition, ownership concentration is also an integral part of privately held

corporations.2In this paper,we study the value of controlling blocks of shares in

public corporations, and thus contribute to our understanding of the costs and

benefits of concentrated ownership. An inherent difficulty in valuing control-

ling blocks of shares is the illiquidity of the market. Theoretically, illiquidity in

∗Albuquerque is from Boston University Questrom School of Business and Cat´

olica-Lisbon

School of Business and Economics. Schroth is from Cass Business School, City University London.

The authors thank Fernando Anjos, Sugato Bhattacharyya, Giovanni Calice, Daniel Carvalho,

Jo˜

ao Cocco, Amy Dittmar, Joost Driessen, Darrell Duffie, Rene Garcia, Nicolae Gˆ

arleanu, Christo-

pher Hennessy, Florencio Lopez de Silanes, Albert Menkveld, Luke Taylor, Christopher Tonetti,

Dimitri Vayanos, the three anonymous referees, the Associate Editor, the Co-Editors, Michael

Roberts and Kenneth Singleton, and the Editor, Bruno Biais, for valuable suggestions that im-

proved the paper. The authors also thank Gonc¸alo Pacheco Pereira for help with data collection.

Albuquerque gratefully acknowledges the financial support from the European Union Seventh

Framework Programme (FP7/2007-2013) under grant agreement PCOFUND-GA-2009-246542 and

from the Foundation for Science and Technology of Portugal.

1Contrary to a long-held belief (e.g., Berle and Means (1932)), Holderness (2009) shows, using

a representative sample of U.S. public firms, that 96% of these firms have blockholders and that

these blockholders own an average of 39% of the common stock. Using a sample of large U.S.

corporations from 1996 to 2001, Dlugosz et al. (2006) find that 75% of all firm-year observations

have blockholders that own at least 10% of the firms’ equity. Holderness, Kroszner, and Sheehan

(1999) report that the mean percentage share ownership of a firm’s officers and directors in 1995

was 21%. See Morck (2007) for evidence outside the United States.

2The Internal Revenue Service estimates that in 2007 the wealth of U.S. investors allocated to

closely held stock (in companies that are not publicly traded) was 62% of the wealth allocated to

publicly traded stock.

DOI: 10.1111/jofi.12207

1405

1406 The Journal of Finance R

the market for controlling blocks is a cost that affects the block value, possibly

in a nonlinear way. Empirically, illiquidity reduces the number of observations

available to the econometrician and constrains the empirical estimation of the

block value. We provide a model of the trading and pricing of controlling blocks

in an illiquid market with search frictions. We argue that block-trading events

convey information that identifies the model parameters and allows the esti-

mation of the value of control.

The model’s main premise is that a controlling blockholder of a public cor-

poration affects the value of the firm’s assets (Holderness and Sheehan (1988),

Barclay and Holderness (1989), and more recently, P´

erez-Gonz´

alez (2004)).

Therefore, given the choice, the controlling blockholder will only sell to a bid-

der who can increase asset value. In addition, we assume that the controlling

blockholder is forced to sell if hit by a liquidity shock, in which case he may

sell to a party that creates less asset value and be paid a fire sale price. The

potential absence of a bidder at any given time further increases the illiquidity

of the block. These frictions give rise to a marketability discount on the value

of the block. In addition, the possibility that the new blockholder may decrease

asset value introduces a discount on the dispersed shares traded in the stock

market. We name this novel effect the illiquidity-spillover discount.

Estimation of the marketability and illiquidity-spillover discounts is noto-

riously difficult because they require a counterfactual analysis: what should

the price be absent search frictions? The structural estimation adopted in the

paper uses the model’s pricing equations to evaluate this counterfactual price.

To do so, the structural estimation must successfully identify the parameters

of these pricing equations. In particular, since the pricing will differ depending

on whether the trade was caused by liquidity shocks, the model must iden-

tify ex post the reasons for trading. These reasons are unobservable to the

econometrician. One contribution of the paper is to show that it is possible to

identify the model’s parameters by using the valuations of two different types

of shareholders during a block trade: the blockholders’ valuation implicit in the

negotiated block price and the dispersed shareholders’ valuation revealed in

the exchange share price.

In the model, a liquidity shock is the realization of a random variable with

a Bernoulli distribution that forces blockholder turnover. Following a liquidity

shock, the block is sold at a fire sale price equal to a fraction of the buyer’s val-

uation. In contrast, the dispersed shareholders only care about the discounted

value of future cash flows under the new blockholder and not the fire sale price.

This price difference allows us to identify fire sale discounts.

In the absence of a liquidity shock, the block changes hands only if a potential

new blockholder arrives and can generate more cash flow. In this case, block

and share prices differ partly because liquidity shocks penalize blockholders

more than dispersed shareholders, who are unaffected by the lower expected

fire sale price in a future sale. In short, our model is able to fully exploit the

data by identifying liquidity shock probabilities not only from the frequency of

trades with negative price reactions, but also from the block versus share price

differences that exist regardless of whether liquidity shocks have occurred.

The Value of Control and the Costs of Illiquidity 1407

We estimate an average probability of getting a liquidity shock within one

year of 20%, and, conditional on a liquidity shock, an average fire sale discount

of 8% of the block value.3The estimated probability of meeting a potential

buyer within one year is 43%. We find that the marketability discount, which

is a nonlinear function of these three estimates, is on average 13% of the block

value, with a standard deviation of 22%. The spillover effect of the block’s

illiquidity on the dispersed shares is on average 2.1% of the share price.4

A selection bias in our estimates may arise if not all liquidity shocks lead to a

fire sale. That is, blockholders may have a reservation value that is determined

by the actions taken to avoid a fire sale, for example, using the block as col-

lateral for a loan. We argue that this selection bias leads to downward-biased

estimates of the probability of a liquidity shock. In addition, our reduced-form

approach to fire sale prices may lead to upward-biased estimates of the fire

sale price. That is, what we call a fire sale price is likely to be the maximum

payout among many alternative ways of reacting to the liquidity shock, which

include, for example, the arrival of a private equity firm supplying liquidity.

Because the marketability and illiquidity-spillover discounts are decreasing in

the fire sale price and increasing in the probability of a liquidity shock, these

biases lead to underestimation of the discounts. Our approach therefore gives

a conservative estimate of the discounts we study.

We allow the probability of a liquidity shock and the fire sale price to vary

across deals as a function of economy-wide and deal-specific determinants of

liquidity in order to match the observed variation in the block and exchange

share prices. Economy-wide determinants of liquidity appear to capture un-

observed variation in the probability of a liquidity shock, whereas firm and

industry characteristics appear to capture unobserved variation in fire sale

values. We find that the probability of a liquidity shock is increasing in the

Fontaine and Garcia (2012) measure of aggregate funding illiquidity, and de-

creasing in GDP growth and the stock market return. However, the probability

of a liquidity shock is high when GDP growth and market returns are high

and the yield curve is steep. Our interpretation is that macroeconomic expan-

sions increase blockholder liquidity via their balance sheet effects but may also

trigger a preference for cash if they bring good investment opportunities when

outside funding is costly.The block’s fire sale value decreases with the degree of

asset specificity of the target firm’s industry and with the target firm’s leverage

relative to that of its industry. The evidence that the state of the aggregate

economy determines firm-specific liquidity complements the work of Chordia,

Roll, and Subrahmanyam (2000) and Bao, Pan, and Wang (2011), who find

commonalities in asset-specific liquidity measures.

3The block fire sale discount estimate is similar to those in other markets: Coval and Stafford

(2007) estimate a 10% discount on stocks that experience price pressure due to mutual fund

outflows, Pulvino (1998) estimates a 14% fire sale discount for aircraft of some airlines, and

Andersen and Nielsen (2013) estimate a 12.5% discount on forced sales in the real estate market.

4The spillover effect is economically significant and equal to five times the size of the mean

equal-weight quoted bid-ask spread on equities (see Bollen, Smith, and Whaley (2004)).