1 of 26

The Use of Escrow Contracts in

Acquisition Agreements

Sanjai Bhagat (Univ. of Colorado - Boulder)

Sandy Klasa (Univ. of Arizona)

Lubomir P. Litov (Univ. of Oklahoma & WFIC, Univ. of Pennsylvania)

June 2023, MARC Bayes Conference, City University London

2 of 26

2

Escrow Contracts in Acquisition Agreements

A form of contingent payment contract, where buyer and seller agree that seller will deposit a fraction of purchase price (usually 12.2% of it) into a treasury account for a certain period of time (usually 17.4 months).

Employed in unlisted firm and subsidiary acquisitions but not used in acquisitions of listed traded targets.

In our sample, escrow accounts used in 52.1% of all unlisted target deals.

More commonly used in (i) stock purchase acquisitions and (ii) acquisitions of stand-alone targets.

Most acquisitions made by publicly traded firms consist of acquisitions of stand-alone private firms or subsidiaries of other firms (e.g., Fuller, Netter, and Stegemoller (2002) and Netter, Stegemoller, and Wintoki (2011)). However, there is only limited empirical evidence on acquisitions of unlisted targets in prior work, which restricts our understanding of these types of acquisitions.

When these contracts are used the bidder places a percentage of the total purchase price in an escrow account, which is held for a negotiated period of time after the completion of the acquisition. Bidders can lay claim on the funds in the escrow account to cover their losses if the sellers fail to meet specific terms of the acquisition agreement or if it is found that negative information about the target was hidden from the bidder prior to the sale. We hypothesize that escrow contracts are an efficient contracting mechanism that can create real value by allowing buyers and sellers to reduce acquisition-related costs and manage the considerable transaction risk inherent in private firm and subsidiary acquisitions.

For a sample of 569 acquisitions of stand-alone private firms and 374 subsidiary acquisitions over the 1994-2009 period, we hand-collect data on escrow contracts from bidder firms’ 8-K, DEF 14A, 10-Q, and 10-K filings. We find that escrow contracts are employed in 52.1% of the acquisitions we study. Also, for the deals with an escrow contract, on average, 12.2% of the sale proceeds are placed in an escrow account and the funds are held in the escrow account for approximately 17.4 months.

Use in public targets: When an escrow contract is included in an acquisition agreement all shareholders need to individually sign a contract with the custodian bank that maintains the escrow account. Hence, escrow contracts are not used in acquisitions of publicly traded firms due to the large number of shareholders who would need to sign these related contracts.

3 of 26

3

Use of Escrow Contract Funds

- EBITDA & Purchase Price Adjustments.

- Working Capital Adjustments (A/R, A/P, Inventory, other current assets).

- Environmental Liabilities.

- Pending Litigation.

- Obligations Related to Collective Bargaining Agreements.

- Unpaid Taxes Due.

- Other Non-Fraudulent Breach of Representations & Warranties.

Escrow Agreement Disputes Resolution

- Binding Arbitration.

- Non-binding Mediation.

Escrow Agreements vs. Transaction Insurance

- Insurance more expense, and limited in coverage.

Institutional Detail

When an escrow contract is included in an acquisition agreement a percentage of the total sale proceeds is placed in an escrow account, which is held for a negotiated period of time after the acquisition. The funds in the escrow account are used to cover bidder losses resulting from seller’s “breaches of representations and warranties made to the bidder in the acquisition agreement” (this needs to be re-phrased). For instance, if after the completion of an acquisition it is found that the sellers misstated in their representations and warranties the target’s earnings before depreciation and amortization, the bidder firm could lay claim to some of the funds in the escrow account. Likewise, a bidder could make a claim on these funds to address purchase price adjustments, if after the acquisition it is determined that the physical condition and value of certain of the target’s assets differ from what was represented by the sellers. Another common reason why a bidder might make a claim on escrow account funds is if the target’s working capital (i.e., accounts receivable, inventory, and other current assets) was misrepresented by the sellers. Also, if the sellers misstate what are the target’s total liabilities, including liabilities such as environmental liabilities, pending litigation, contractual obligations related to collective bargaining agreements, or unpaid taxes due, the bidder could make a claim on some of the escrow account funds.

Upon negotiating what is the percentage of the sale proceeds placed in the escrow account and how long the funds are kept in the account, the bidder and sellers also decide on an escrow agent and an arbitrator to resolve disputes related to bidder claims on the escrow account funds. The escrow agent is often a division of a bank holding company, such as the escrow services division of J.P. Morgan Chase. The funds placed in the escrow account are held by the escrow agent until the remaining funds are to be disbursed back to the sellers of the target, as stipulated in the escrow contract. In return for being the custodian of the funds, the escrow agent typically receives a fixed fee and earns some interest on the funds with the remainder of the interest earned on the funds going to the sellers of the target.

Even though using an escrow contract can significantly reduce transaction-related litigation after deal completion, disputes may arise as to the use of the escrow account funds. When such disputes occur, parties commonly rely on the provisions of the escrow contract for arbitration. For that purpose, arbitrators used in conjunction with escrow contracts are typically hired from organizations that specialize in alternative dispute resolution (i.e., alternative to litigation in court). Examples of these organizations include the American Arbitration Association and the Judicial Arbitration and Mediation Services. These organizations typically hire arbitrators who are retired judges, attorneys, or business executives with relevant experience and expertise. In return for their services, arbitrators typically earn a fixed fee and/or a variable fee that is a percentage of the dispute value. Arbitrators’ decisions relating to bidder claims on escrow account funds can be binding or non-binding, and that is written up in the escrow contract. In the case of non-binding arbitration, the arbitrator’s decision could be challenged in court, which would lead to additional legal costs and to information about the dispute becoming public knowledge. Therefore, using binding arbitration is typically the more efficient and commonly observed approach.

- An alternative to using an escrow contract to manage acquisition-related transaction risk is to purchase insurance to manage this risk from an insurance brokerage firm such as Marsh & McLennan Company. However, the costs for a bidder to purchase such insurance are usually considerably higher than would be the costs involved in establishing an escrow contract. Moreover, coverage by such insurance policies is usually very limited in scope when compared to uses of escrow funds as written in the escrow contract.

- We note that in cases where the method of payment is a combination of cash and bidder stock, cash is placed in the escrow account. However, if the method of payment consists only of bidder stock then the funds placed in the escrow account are in the form of bidder stock and no interest is earned on these funds.

- The bidder and seller could also stipulate in the escrow contract that a mediator will be used to help resolve disputes. However, mediators’ suggested solutions to a dispute are non-binding, and as such, in most cases using an arbitrator who makes binding decisions is anticipated to be a more efficient way to resolve escrow contract disputes.

The use is as follows: claims to cover

breach of reps and warranties in the contract,
addressing purchase price adjustments;
working capital changes (accounts receivable);
unpaid due tax;
environmental liabilities;
pending litigation.

4 of 26

4

Preliminary Sample Statistics

Variable	All Sample (943 obs.)
Percent of Deals with an Escrow Agreement	52.1%

Variable	Escrow Contract (491 obs.)
Percent Deposited Through an Escrow Agreement	12.2%
Average Escrow Agreement Size (US$ m)	$11.7
Average Escrow Agreement Duration (Months)	17.4

All Deals Sample

Deal with Escrow Contract

Percent of deals with:	All	Sample with Escrow Contracts	Sample without Escrow Contracts
% Stock Purchases	73.7%	80.0%	66.8%
% Asset Purchases	26.3%	20.0%	33.2%
% Stand-alone Targets	60.30%	75.60%	43.80%
% Subsidiary Targets	39.70%	24.40%	56.20%

5 of 26

5

Main Hypothesis

The use of escrow contracts in unlisted target acquisitions is an efficient contracting mechanism that:

Facilitates completion of these acquisitions by allowing buyers & sellers to manage acquisition-related transaction risks, and

2) Allow parties to overcome information asymmetry problems.

Hypotheses

6 of 26

6

Several implications of the main research hypothesis follow through.

Use of Escrow Contracts - Transaction Risk Predictions

Stand-alone targets are likelier to use escrow contract.
Subsidiary targets less likely to use escrow contract.
Larger targets (relative to acquirer) are likelier to use escrow contract.
Targets with dominant shareholders are likelier to use escrow contract (to manage joint & several liability if post-acquisition breach of reps. & warranties).
Stock purchases (as opposed to asset purchases) are likelier to use escrow contract.

Hypotheses

7 of 26

7

Use of Escrow Contracts - Informational Asymmetry Risk Predictions

Different industry acquisitions likelier to use escrow contract.
Targets w/ higher degree of earnings management likelier to use escrow contract.
Targets in industries w/ higher earnings volatility likelier to use escrow contract.
Targets in industries w/ lower analyst coverage likelier to use escrow contract.
Distressed targets (i.e., w/ lower interest coverage) likelier to use escrow contract.

Use of Escrow Contracts – Other Risks Predictions

Escrow contracts are used if reverse insurance in the form of liability cap limitations (or caps) is included (caps are on average 2.4 times escrow size).

Hypotheses

Escrow contract use more likely when information asymmetry in the economy is higher (i.e., when C&I spread is higher).

A side note, escrow contracts are less likely to be used if an installment note sale is used.

We investigate whether the likelihood that an escrow contract is employed in the context of an acquisition is greater when it is more important for the bidder to manage acquisition-related transaction risk. We first compare the use of escrow contracts in subsidiary versus private stand-alone firm acquisitions. In the case of subsidiary acquisitions, a bidder would have legal recourse post-deal closure against both the parent firm that sold the subsidiary and the principal shareholders of the parent firm, while for private firm acquisitions the bidder would only have recourse against the principal shareholders of the target. Consequently, we expect that bidder acquisition-related transaction risk is higher in the case of private firm acquisitions than subsidiary acquisitions, and that as a result, escrow contracts are used more often in private firm acquisitions. Our findings are consistent with this prediction. We document that escrow contracts are used in 65% of private firm acquisitions, but that these contracts are used in only 32% of subsidiary acquisitions. Further, in multivariate tests we show that whether an acquisition is a private firm acquisition positively impacts the likelihood that an escrow contract is used.

When the relative size of the target to the bidder is greater it should be more important for the bidder to manage acquisition-related transaction risk. Thus, in these instances the use of an escrow contract should be more prevalent. Supporting this prediction, we find that the relative size of the target to the bidder has a positive effect on the likelihood that an escrow contract is employed in the context of an acquisition.

Because when there is more information asymmetry about the target’s value the bidder faces greater transaction risk, in these instances the benefits to using an escrow contract should be larger. We consider five cases in which information asymmetry about a target’s value is likely greater. First, we assume there is more information asymmetry about a target’s value if it operates in an industry with higher earnings volatility. Second, we expect that if there are fewer distinct analysts covering the publicly traded firms in a target’s industry this makes it more difficult to value unlisted targets in the industry, which leads to greater information asymmetry about these targets. Third, we consider if the target operates in a different industry than does the bidder, assuming that if so, then information asymmetry problems are larger. Fourth, we look to a target’s total accruals, as measured by the difference between its earnings and its free cash flows and follow prior work (e.g., Dechow and Dichev (2002)) that argues that when a firm’s total accruals are greater it is more difficult for investors to infer the firm’s financial performance and its value from its earnings. Finally, we expect that there is greater information asymmetry about the value of financially distressed targets. Here, we use a target’s interest coverage ratio to proxy for if it is financially distressed. Consistent with expectations, we find that whether an escrow contract is used in the context of the acquisition of an unlisted target is positively associated with: (1) earnings volatility in the target’s industry, (2) if there is a smaller number of analysts covering the target’s industry, (3) if the target operates in a different industry than does the bidder, (4) the target’s total accruals, and (5) if a target’s interest coverage ratio is low.

If a target has a dominant shareholder, defined as a shareholder who owns at least 20 percent of the target’s shares but not all of its shares, an escrow contract can be particularly useful to manage this shareholder’s acquisition-related transaction risk. If such a contract is in place all target shareholders would bear pro rata costs of bidder recourse actions subsequent to an acquisition. In contrast, if such a contract is not in place bidder recourse actions subsequent to an acquisition could result in the target’s dominant shareholder being held liable and sued by the bidder. Supporting the proposition that an important motive for escrow contracts is that they help reduce the transaction risk of a target’s dominant shareholder, we find that the presence of a target dominant shareholder positively impacts the likelihood that an escrow contract is used.

We also investigate whether the use of a cap on the amount a bidder could sue the sellers for subsequent to an acquisition is associated with the use of an escrow contract. If a bidder agrees to such a cap this reduces the sellers’ acquisition-related transaction risk, and as a result, this should make it easier to persuade the sellers to agree to include an escrow contract in the acquisition agreement. Supporting this notion, we show that the use of a cap on how much future compensation the bidder could seek from the sellers positively impacts whether an escrow contract is used in conjunction with the acquisition.

In such cases, the dominant shareholder could then try to sue the smaller shareholders to recover some of their share of the sale proceeds. However, given that these recourse lawsuits would be costly, the use of an escrow contract would be a more efficient way for this shareholder to manage acquisition-related transaction risk.

8 of 26

8

Escrow agreements are more common when transaction risk is high, i.e., for:

Stand-alone private targets, targets with dominant shareholders, larger targets (relative to acquirer).
Escrow contract use shortens time-to-complete transaction by 35.5-51.0%.

Escrow agreements are more common if information asymmetry risk is high, i.e., for:

Targets in cross-industry acquisitions, targets w/ high accruals, targets in industries w/ high earnings volatility or low analyst coverage, targets that are financially constrained.

Escrow agreements are more common when other risks may be present, i.e.:

If no reverse insurance in form of liability cap limitations (or caps) is included.

Main Findings

Escrow contract use shortens time-to-complete transaction from average of 57.5 days to 37.1 days (OLS).

In acquisitions of private firm and subsidiary targets a bidder’s due diligence costs are often large relative to the value of the target because of asymmetric information problems. However, the use of an escrow contract is expected to reduce a bidder’s need to incur significant due diligence costs in these deals. We run tests that provide insights on the extent to which an escrow contract creates real value by reducing a bidder’s due diligence costs. To do so, we proxy for these costs with the length of the period between the announcement of a preliminary acquisition agreement and the completion of the deal, assuming that when this period is longer a bidder conducts more due diligence on the target. We focus on the period between these two dates because prior to the signing of the preliminary acquisition agreement the sellers of a target will typically only provide a bidder with a limited amount of private information about the target and thus it would not be possible for the bidder to conduct major due diligence on the target. On the other hand, after this agreement is signed the target’s sellers are contractually obligated to provide the bidder with broad access to private information about the target, and the bidder can then begin extensive due diligence on the target, such as verifying the accuracy of the target’s accounting information, the physical condition of the target’s assets, or whether the sellers of the target have appropriately represented the target’s total liabilities.

We find evidence consistent with the use of escrow contracts leading bidders to reduce the amount of due diligence they conduct on an unlisted target. Specifically, we document that even after controlling for the potential endogeneity of having an escrow contract in place, that the use of such a contract is associated with about a 50% decrease in the length of the period between the announcement of a preliminary acquisition agreement and the completion of the deal.

Wangerin (2010) validates the use of the length of the period between the announcement of a preliminary acquisition agreement and the completion of the deal as a proxy for a bidder’s due diligence costs. Also, he shows that the length of this period is positively associated with information asymmetry about a target’s value and acquisition-related litigation risk.

9 of 26

9

Escrow contracts are associated with significantly smaller target discount:

- Adoption of escrow agreement increases valuation of a private target (price-to-sales multiple compared to similar public target) by 3.5-6.1%.

- Valuation impact stronger for stand-alone targets (7.4-9.1%) and weaker for subsidiary targets (0.8-2.2%).

Escrow contracts benefit bidder:

- For every dollar in escrow contract deposit, market capitalization to bidder shareholders is increased by 89 cents.

- Bidder benefits from reduced losses (due to unrecorded or uncovered target liabilities), reduced information gathering (due diligence) costs, and reduced litigation costs (in case of dispute with target after closing).

Main Findings

Subsequently, we determine the effect of the use of an escrow contract on the total amount of the proceeds received by the sellers of an unlisted target after accounting for the portion of the funds in the escrow account predicted to be kept by the bidder. As discussed above, the use of an escrow contract is expected to result in real value creation because it ought to reduce a bidder’s need to incur large due diligence costs. If the bidder agrees to pass on to the sellers a portion of this created value then we should observe that the use of an escrow contract positively impacts the total amount of the proceeds received by the sellers of an unlisted target. Given the evidence in Officer (2007) that in stand-alone private firm and subsidiary acquisitions targets are typically sold at important discounts relative to the prices paid for comparable public targets, we examine if the use of an escrow contract reduces this price discount. After controlling for the potential endogeneity of having an escrow contract in place and the potential costs to the sellers if the bidder ends up receiving some of the funds in the escrow account, we estimate that, on average, the use of an escrow contract is associated with approximately a 6% increase in the total proceeds received by the sellers of an unlisted target. This effect is more pronounced in instances when the use of an escrow contract is expected to reduce bidder due diligence costs to a larger degree, and consequently, the real value created by the use of an escrow contract would be larger. Specifically, we find that the extent to which the use of an escrow contract reduces the discount in the final price paid for an unlisted target is greater when a target is a private stand-alone firm, the relative size of the target to the bidder is larger, or there is likely more information asymmetry about the target’s value.

Lastly, we examine whether the use of an escrow contract in the acquisition of an unlisted target leads to more positive market reactions for bidders to the announcement of the acquisition deal. This should be the case if not all of the cost savings for a bidder from using an escrow contract are passed on to the target in the form of a higher acquisition price. We find that bidder acquisition announcement returns are positively associated with whether an escrow contract is used in conjunction with the acquisition of an unlisted target, and also show that this result is robust to controlling for the potential endogeneity of having an escrow contract in place. Further, we document that this effect is more pronounced when the reduction in bidder due diligence costs from using an escrow contract is anticipated to be greater, as proxied for by a target being a stand-alone private firm, a higher value for the relative size of the target to the bidder, or the existence of greater information asymmetry about the value of the target.

10 of 26

10

Data

Use Business Valuation Resources LLC data.

- Information on private target financials from 8K, DEF 14A, 10Q, & 10K.

Collect data on 943 acquisitions:

- Involve private target & public acquirer.

- Target value above $25 million.

- Completed 1994-2009.

- Available information on target financials & consideration structure.

We hand collect data on the presence of escrow contracts, caps, & earn outs.

Collect data on size, maturity & type of consideration offered in escrow contract.

Other: exclusive access to J.P. Morgan Chase escrow services client proprietary data.

11 of 26

11

Univariate Sample Statistics

Table I. Panel A. Escrow Contracts and Target Price Discount Characteristics.

Variable	All Sample (943 obs.)
Percent of Deals with an Escrow Agreement	52.1%
Average Time to Completion	57.5 days

Variable	Escrow Contract (491 obs.)	Stand-alone Targets (371 obs.)	Subsidiary Targets (120 obs.)
% Deposited Through Escrow Contract	12.2%	13.0%	8.9%
Escrow Contract Size (US$ m)	$11.7	$11.9	$11.0
Escrow Contract Duration (Months)	17.4	17.5	17.1

All Deals Sample

Deal with Escrow Contract

Unlisted Target Price Discount (Means)

	Discount
All Targets	-0.243
Stand-alone Private Targets	-0.171
Subsidiary Targets	-0.333
Targets With Escrow Contract	-0.196
Targets Without Escrow Contract	-0.289

12 of 26

12

Percent of deals with:	All	Sample with Escrow Contracts	Sample without Escrow Contracts
% with Dominant Owner	36.70%	48.30%	24.10%
% with Liability Cap	73.7%	85.3%	61.1%
% with Stock Payment Only	9.0%	11.2%	6.6%
% with Cash Payment Only	36.7%	31.8%	42.0%
% with Earnout Contract	12.2%	13.8%	10.4%
% with Installment Note	11.7%	8.8%	14.8%

Use of Escrow Contracts and Other Deal Characteristics

Table II. Panel A. Deal Characteristics.

Table II, Panel B. Sub-sample Comparisons.

	Escrow Deals with row characteristics	Escrow Deals without row characteristics
Stock Purchase	60.5%	43.5%
Stand-Alone Target	65.2%	32.1%
Dominant Target Owner	68.5%	42.5%
Use of Liability Cap	60.3%	29.1%
Payment is Stock Only	64.7%	50.8%
Payment is Cash Only	45.1%	56.1%
Use of Earnout Contract	59.1%	52.1%

Discuss briefly:

Stand alone vs. private

Asset vs. Stock

Panel A documents tabulations within the sets of escrow and no-escrow deals.

Panel B documents tabulations within the sets of stock purchases, asset purchases, stand alone targets, etc.

Panel A presents tabulations of what percent of escrow deals have the row characteristic.

Panel B presents tabulations of what percent of the row characteristic group has an escrow.

Panel B of Table II compares the usage of an escrow contract between deals with different characteristics. Consistent with the Panel A results, this panel shows that the percentage of deals using an escrow contract is higher for deals that are stock purchase transactions rather than asset sales (60.5% versus 43.5%), deals that are the acquisition of a private stand-alone firm rather than a subsidiary target (65.2% versus 32.1%), deals in which there is a dominant target shareholder (68.5% versus 42.5%), deals using a liability cap (60.3% versus 29.1%), or deals in which the method of payment consists only of stock (64.7% versus 50.8%). Also, Panel B documents that the percentage of deals using an escrow contract is lower when the method of payment is only cash (45.1% versus 56.1%), but that this percentage does not significantly differ between deals that use or do not use an earnout contract.

13 of 26

13

Determinants of Escrow Contract Use in Unlisted Target Acquisitions (Table IV)

(Probit Models)	Use of Escrow Agreement
Variables:	All Targets	Stand-alone Targets	Subsidiary Targets
Target is Stand-alone Private Firm	0.223^***
(t-stat)	(5.68)
Indicator for Asset Sale	-0.017	-0.03	-0.003
	(0.41)	(0.47)	(0.05)
Relative Size of Target to Acquirer	0.019^*	0.021^**	0.008
	(1.66)	(2.53)	(1.18)
Top Quintile Earnings Volatility Target Industry	0.009^**	0.086^**	0.108
Indicator	(2.04)	(2.27)	(1.20)
Low Analyst Coverage Target Industry	0.038^*	0.043^*	0.112^**
Indicator	(1.79)	(1.87)	(2.12)
Target & Acquirer are in Different Industry	0.016^**	0.041^***	0.025
	(2.48)	(2.95)	(0.43)
Target Accruals	0.542^**	0.515^*	0.799^*
	(2.15)	(1.71)	(1.82)
Target Interest Coverage	-0.079^*	-0.069^**	-0.031
	(1.88)	(2.00)	(1.58)
Indicator for Target Dominant Shareholder	0.161^***	0.125^***	0.218^***
	(4.51)	(2.95)	(2.75)
Indicator for Use of Liability Caps	0.267^***	0.312^***	0.228^***
	(6.97)	(5.53)	(3.70)
C&I Loan Rate Spread	-0.005	-0.023	-0.073
	(0.10)	(0.36)	(0.73)
Ln (Deal Value)	-0.046^***	-0.024	-0.073^***
	(2.78)	(1.06)	(2.72)
% Deals Target Industry Using Escrow Contract	0.208^*	0.426^**	0.046
	(1.68)	(2.23)	(1.21)
% Bidder Use of Hybrid Method of Payment	0.012^*	0.010^**	0.047^**
	(1.92)	(2.08)	(2.48)

(continued)

Redo the order to be the set of the three sets of explanations: transaction risk, asymmetric information, other.

Consistent with expectations, we find that whether an escrow contract is used in the context of the acquisition of an unlisted target is positively associated with: (1) earnings volatility in the target’s industry, (2) if there is a smaller number of analysts covering the target’s industry, (3) if the target operates in a different industry than does the bidder, (4) the target’s total accruals, and (5) if a target’s interest coverage ratio is low.

It is also possible that within deals that use an escrow contract the fraction of the sale proceeds held in the escrow account is associated with buyer or seller acquisition-related transaction risk. To examine this issue, for the subsample of 491 deals in which an escrow contract is used, we re-estimated the first model from Table IV using the OLS procedure and replacing the dependent variable in this model with the fraction of the sale proceeds held in the escrow account. We find that the majority of the proxy variables for bidder and seller acquisition-related transaction risk retain significant coefficients with expected signs. Specifically, the following proxy variables for this risk retain statistically significant coefficients with predicted signs: whether the target is a stand-alone private firm, target relative size, the low target industry analyst coverage indicator, whether the target and the acquirer are in different industries, target accruals, and the indicator for if a liability cap is included in the acquisition agreement. Further, although in the first model of Table IV we find an insignificant coefficient on the indicator variable for if the acquisition is considered to be an asset sale, when we re-estimate this model replacing the dependent variable with the fraction of the sale proceeds held in the escrow account, we find, as predicted, a statistically significant positive coefficient on this indicator variable. Overall, these results are consistent with the notion that the larger is acquisition-related transaction risk the greater is the fraction of the sale proceeds placed in the escrow account.

14 of 26

14

Endogeneity & Instrumental Variables Choice

Adoption of escrow contract is endogenous w.r.t. valuation.

Need instruments that are both (i) relevant & (ii) exclusive to determine choice of escrow adoption.

Two instruments:

- Propensity of peer target industry group to use such contracts.

- Bidder reputation to use hybrid securities (including escrow contracts).

Verify validity of such instruments through series of tests:

- Overidentification test (Sargan stat).

- Underidentification tests (partial R-squared, F-stat excluded instruments).

These come in response to endogeneity concerns in multivariate analysis.

The main issue here is that the choice of an escrow contract in the agreement is non-random, and such in a predictable way. To the extent that one reason to have an escrow is information asymmetry, and to the extent that information asymmetry also influences target valuation, one can argue that we are not properly addressing the endogenous choice of having an escrow contract. To amend it, we use two instruments.

The first is the previous use in the target industry of escrow contracts, which by hearsay among peers may be used by the target in point. The second instrument we use is the reputation by buyer for prior use of hybrid securities. Some serial acquirers, such as Emerson, would use the escrow agreements extensively, but rarely file claims to these subsequent to closing. By establishing a track record (a reputation) of use of escrow agreements (as part of the stock purchase, asset purchase or statutory merger) which are rarely used by buyer to file claims subsequent to closing, a buyer gains advantage in the process of negotiations with the seller . Granted, this however is a second-order effect, and we do not test it per se. These instruments are likely to influence the choice of having an escrow, but not likely to influence the valuation of the target.

Instruments:

The first variable is the percentage of deals for unlisted targets in a target’s Fama-French 49 industry that used an escrow contract during the previous year (calculated using data from our sample). Here, we predict that the higher is this percentage the more likely it is that an escrow contract is included in the acquisition agreement.

The second variable is an indicator for if during the prior year the bidder paid for an acquisition using a hybrid method of payment, defined by SDC as a mix of cash and securities in the bidder firm. We expect that bidders that previously used a hybrid method of payment to pay for an acquisition are likely to offer more complex types of payment when making an acquisition, such as including a payment that would be placed in an escrow account. Thus, we predict this indicator is positively associated with the likelihood that an escrow contract is included in the acquisition agreement.

To determine the validity of our 2SLS analyses, we perform diagnostic tests to examine the suitability (i.e., relevance) of the instruments used for these analyses and the appropriateness (i.e., exclusivity) of using an instrumental variables approach.

First, the results of an F-test of excluded instruments and a partial R-squared test for the first-stage regressions indicate that the instruments in our first-stage models are jointly significant in explaining if an escrow contract is included in a given acquisition agreement.

Second, the results of the Anderson (1951) canonical correlation likelihood ratio test reject the hypothesis that our 2SLS analyses suffer from weak instrument problems.

Next, in order to verify the exclusivity (i.e., if the effect of our instruments onto the endogenous channel is confined to the first stage equation alone) of our instruments, we ran a Sargan test of over-identifying restrictions and found that the second-stage regressions do not suffer from over-identification problems, meaning that our instruments are exclusive. Based on these tests, we conclude that our instruments are both (i) relevant and (ii) exclusive. The

F-test is based on the F-statistic of excluded instruments and tests the joint significance of all excluded instruments in the first-stage regressions. The partial R-squared is the fraction of the variation of the instrumented variable explained by the instruments, net of their effect through the exogenous variables.

The Anderson (1951) canonical correlation likelihood ratio statistic examines the null hypothesis of under-identification using the canonical correlation between the regressors and their instruments.

Sargan Test

The test of over-identifying restrictions (i.e., a Sargan test) tests the joint null hypothesis that the instruments are uncorrelated with the error term and are hence correctly excluded from the second-stage equations.

15 of 26

15

Excluded Instruments:	All Targets	Stand-alone Targets	Subsidiary Targets

Percent of Deals in Target Industry that Use Escrow Agreement in the Previous Year	0.208^*	0.426^**	0.046
(t-stat)	(1.68)	(2.23)	(1.21)
Indicator Previous Use of Hybrid Securities by Acquirer	0.012^*	0.010^**	0.047^**
	(1.92)	(2.08)	(2.48)

Observations	837	520	317
Pseudo R-squared	0.205	0.278	0.358

Escrow Use

Reputation

proxy

Industry

proxy

Determinants of Escrow Contract Use in Unlisted Target Acquisitions (Table IV)

The first variable is the percentage of deals for unlisted targets in a target’s Fama-French 49 industry that used an escrow contract during the previous year (calculated using data from our sample). Here, we predict that the higher is this percentage the more likely it is that an escrow contract is included in the acquisition agreement.

The second variable is an indicator for if during the prior year the bidder paid for an acquisition using a hybrid method of payment, defined by SDC as a mix of cash and securities in the bidder firm. We expect that bidders that previously used a hybrid method of payment to pay for an acquisition are likely to offer more complex types of payment when making an acquisition, such as including a payment that would be placed in an escrow account. Thus, we predict this indicator is positively associated with the likelihood that an escrow contract is included in the acquisition agreement.

16 of 26

16

Economic Significance of Estimates in Determinants Model

Model of Escrow Contract Use Determinants

Independent Variables:	∆ in Probability
Main Determinants
Indicator for Use of Caps	26.70%
Target is Stand-alone Private Firm	22.30%
Indicator for Presence of Dominant Shareholder	16.10%
Low Analyst Coverage Target Industry Indicator	3.80%
Target Accruals^*	3.69%
Target & Acquirer are in Different Industry	1.60%
Top Quintile Earnings Volatility Target Industry Indicator	0.90%
Target Relative Value^*	0.64%
Target Interest Coverage^*	-0.22%
C&I Spread^*	-0.93%
Indicator for Asset Sale	-1.70%
Ln (Target Value)^*	-7.05%
Instrumental Variables
% of Deals in Target Industry that Use Escrow Prior Year	4.65%
Indicator if Bidder Uses a Hybrid Payment in Prior Year	1.20%

17 of 26

17

Due Diligence Costs Analysis

Bidders use escrow accounts to diminish due diligence costs.

No direct proxy for such costs.

We proxy for due diligence costs by time-to-deal completion.

- Due diligence costs proportionate to time-to-deal completion.

We find supporting results using litigation propensity.

- Due diligence costs may be direct or indirect.

- Inexpensive due diligence may result in future losses & litigation.

- We can use litigation as another proxy for such costs.

Our findings support hypothesis that escrow contracts diminish due diligence costs.

18 of 26

18

Time-to-Deal Completion Analysis (Table V)

	All Targets		Stand-alone Targets		Subsidiary Targets
Independent Variables:	OLS	2SLS	OLS	2SLS	OLS	2SLS
Escrow Agreement Indicator	-20.385^***		-18.714^*		-24.175^**
(t-stat)	(2.68)		(1.71)		(1.99)
Instrumented Escrow Agreement Indicator		-53.841^**		-40.173^**		-21.001^**
		(2.57)		(2.58)		(2.36)
Target is a Stand-alone Private Firm	0.740	15.811
	(0.08)	(0.74)
Acquisition is Considered an Asset Sale	0.835	0.874	0.097	2.414	10.601	8.838
	(0.08)	(0.08)	(0.01)	(0.15)	(0.91)	(0.70)
Relative Size of Target to Acquirer	3.286	5.677	3.997	6.273	-0.133	-3.071
	(0.55)	(0.91)	(0.47)	(0.72)	(0.01)	(0.31)
Top Quintile Earnings Volatility Target Industry	6.866^**	8.904^**	15.416^*	18.900^**	-11.318	-14.321
Indicator	(2.45)	(2.58)	(1.76)	(1.97)	(0.78)	(0.95)
Low Analyst Coverage Target Industry Indicator	23.361^**	28.635^***	36.011^***	39.560^***	0.768	2.102
	(2.32)	(2.77)	(2.87)	(2.75)	(0.05)	(0.12)
Target & Acquirer are in Different Industries	3.714^***	0.803	-0.801	0.172	10.641^*	1.210
	(2.64)	(1.13)	(0.13)	(0.03)	(1.82)	(0.70)
Target Accruals	12.088	25.819	-23.223	-64.083	61.004	31.847
	(0.39)	(0.41)	(0.57)	(1.04)	(1.04)	(0.47)
Target Interest Coverage (scaled by 1,000)	35.997	40.263	65.249	70.500	0.623	6.413
	(0.91)	(0.89)	(0.91)	(0.93)	(0.06)	(0.42)
Indicator for Target Dominant Shareholder	-3.554	-2.950	-11.566	-16.647	7.439	-16.441
	(0.54)	(0.18)	(1.38)	(1.19)	(0.53)	(0.90)
Indicator for Use of Liability Cap in Acquisition	8.968	18.447	11.587	-10.284	8.842	0.454
	(1.17)	(0.75)	(1.13)	(0.45)	(0.78)	(0.02)
Commercial & Industrial Loan Rate Spread	6.572	9.397	27.010^**	31.554^**	-14.291	-7.426
	(0.73)	(0.92)	(2.13)	(2.14)	(0.98)	(0.42)
Ln (Deal Value)	11.238^**	8.420	5.916	6.505	21.079^***	23.665^***
	(2.25)	(1.21)	(1.19)	(1.36)	(3.19)	(2.86)
Observations	884	837	536	520	348	317
Adjusted R-squared	0.086	0.090	0.133	0.132	0.034	0.027

We find evidence consistent with the use of escrow contracts leading bidders to reduce the amount of due diligence they conduct on an unlisted target. Specifically, we document that even after controlling for the potential endogeneity of having an escrow contract in place, that the use of such a contract is associated with about a 50% decrease in the length of the period between the announcement of a preliminary acquisition agreement and the completion of the deal.

Average Times to Completion

All 57.5

Stand-alone 48.2

Subsidiary 71.6

Anticipated probability of escrow

All .545

Stand-alone .626

Subsidiary .415

Economic Effects

All 35.5 – 51%

Stand-alone 38.8 – 52.2%

Subsidiary 12.2 – 33.8%

The first model in Table V shows that, as expected, the time-to-completion of an acquisition deal is negatively associated with whether an escrow contract is included in the acquisition agreement. The coefficient estimate on the escrow agreement indicator variable indicates that the use of an escrow contract is associated with a 20.4 day reduction in the time-to-completion of an acquisition deal. Relative to the mean value for our sample for the time-to-completion for a deal of 57.5 days, this represents a 35.5% decrease in the typical time-to-completion of a deal.

In the second model we attempt to control for the potential endogeneity of whether an escrow contract is used in a particular acquisition deal by replacing the escrow agreement indicator variable with the instrumented escrow agreement indicator calculated from the first model in Table IV. Here also, we document a negative effect of the use of an escrow contract on the time-to-completion of a deal. Multiplying the coefficient estimate of -53.84 with the mean value for the instrumented escrow indicator variable of 0.544, we estimate that the results for the second model imply that the use of an escrow contract is associated with a 29.3 day decrease in the time-to-completion of a deal. This represents a 51.0% reduction in the time-to-completion of a deal.

Fama-French 49 industry (of the acquirer) fixed effects and year fixed effects are included in all the models.

In acquisitions of private firm and subsidiary targets a bidder’s due diligence costs are often large relative to the value of the target because of asymmetric information problems. However, the use of an escrow contract is expected to reduce a bidder’s need to incur significant due diligence costs in these deals. We run tests that provide insights on the extent to which an escrow contract creates real value by reducing a bidder’s due diligence costs. To do so, we proxy for these costs with the length of the period between the announcement of a preliminary acquisition agreement and the completion of the deal, assuming that when this period is longer a bidder conducts more due diligence on the target. We focus on the period between these two dates because prior to the signing of the preliminary acquisition agreement the sellers of a target will typically only provide a bidder with a limited amount of private information about the target and thus it would not be possible for the bidder to conduct major due diligence on the target. On the other hand, after this agreement is signed the target’s sellers are contractually obligated to provide the bidder with broad access to private information about the target, and the bidder can then begin extensive due diligence on the target, such as verifying the accuracy of the target’s accounting information, the physical condition of the target’s assets, or whether the sellers of the target have appropriately represented the target’s total liabilities.

19 of 26

19

Unlisted Target Discount

We select to study valuation impact for target based on deal-value-to-sales multiple.

- Officer (2007) uses deal-value-to-EBITDA multiples.

- We do not use such multiples because:

* EBITDA could be <0 (for public bidder or private target).

* EBITDA defined discretionary across private targets.

We calculate discount as ratio of private target multiple to similar public target multiple (net of one).

Matching of public to private target counterparts:

- Same industry (2-digit SIC).

- Within 20% difference ideal size.

- Within 36-months window centered on acquisition announcement date.

To calculate the discount in the price of an unlisted target we use an approach similar to that in Officer (2007).

The discount in the price paid for an unlisted target is calculated as the percent difference between the deal-value-to-sales multiple for the unlisted target and the median deal-value-to-sales multiple for the portfolio of comparable public target acquisitions.

We match each private firm or subsidiary target to a portfolio of comparable acquisitions of publicly traded targets that are identified from the SDC database. The acquisitions of publicly traded targets are permitted to enter multiple portfolios matched to unlisted targets. The comparable acquisitions of publicly traded targets are for targets in the same two-digit SIC code industry as the unlisted target in our sample for which the deal value is within 20% of the deal value of the unlisted target. We also require that the comparable transactions are announced within the 36-month window centered on the announcement of the acquisition of the unlisted target. To calculate the unlisted target discount, we use the deal-value-to-sales multiple. We do not use the deal-value-to-EBITDA multiple because it is negative for 32% of private firm targets and 31% of subsidiary targets in our sample. The discount in the price paid for an unlisted target is calculated as the percent difference between the deal-value-to-sales multiple for the unlisted target and the median deal-value-to-sales multiple for the portfolio of comparable public target acquisitions.

20 of 26

20

Analysis of Unlisted Target Discount

Table VI. Panel A.

	All Targets		Stand-alone Targets		Subsidiary Targets
	OLS	2SLS	OLS	2SLS	OLS	2SLS
Escrow Agreement Indicator	0.084^**		0.126^**		0.044^*
(t-stat)	(2.30)		(2.35)		(1.65)
Instrumented Escrow Agreement Indicator		0.202^***		0.228^***		0.139^***
		(3.16)		(3.28)		(2.73)
Target is a Stand-alone Private Firm	0.159^***	0.197^***
	(4.00)	(3.82)
Acquisition is Considered Asset Sale	0.025	0.029	0.121^*	0.141^**	-0.019	-0.009
	(0.63)	(0.69)	(1.94)	(2.16)	(0.34)	(0.16)
Relative Size of Target to Acquirer	-0.051^***	-0.04^**	0.012	0.019	-0.093^***	-0.069^*
	(2.72)	(2.26)	(0.18)	(0.29)	(2.84)	(1.79)
Target & Acquirer in Different Industries	0.070^*	0.061	0.050	0.039	0.088^**	0.056^*
	(1.92)	(1.60)	(0.95)	(0.71)	(2.47)	(1.86)
Top Quintile Earnings Volatility Target Industry	0.121^**	0.133^**	0.187^**	0.177^**	0.060	0.062
	(1.99)	(2.15)	(2.10)	(1.97)	(0.67)	(0.62)
Low Target Industry Analyst Coverage	-0.133^*	-0.130^*	-0.066	-0.096	-0.203^*	-0.222^*
	(1.93)	(1.81)	(0.70)	(1.00)	(1.84)	(1.75)
Method of Payment is Only Cash	-0.011	-0.014	-0.038	-0.039	0.028	0.020
	(0.29)	(0.38)	(1.63)	(1.61)	(0.52)	(0.35)
Commercial & Industrial Loan Rate Spread	-0.019	-0.019	-0.019	-0.046	-0.013	0.019
	(0.33)	(0.33)	(0.24)	(0.56)	(0.15)	(0.19)
Ln (Deal Value)	0.042^**	0.039^*	0.044	0.039	0.046^*	0.060^*
	(2.23)	(1.85)	(1.52)	(1.32)	(1.76)	(1.88)

Observations	889	825	530	508	359	317
Adjusted R-squared	0.218	0.226	0.179	0.191	0.225	0.264

Effect is stronger for stand-alone targets.

Fama-French 49 industry (of the acquirer) fixed effects and year fixed effects are included in all the models.

Table VI, Panel A provides evidence on whether the use of an escrow contract is associated with a reduction in the discount in the price paid for an unlisted target. The first model in this table regresses the discount in the deal-value-to-sales multiple of an unlisted target relative to the deal-value-to-sales multiple of comparable publicly traded targets on an indicator variable for whether an escrow contract is used. The second model in this table accounts for the potential endogeneity in the decision to use an escrow contract by replacing the indicator variable for whether an escrow contract is used with the fitted value calculated from the first model in Table IV.

In the Table VI, Panel A models we control for the following deal characteristics which are related to bidder acquisition-related transaction risk, whether a target is a stand-alone private firm, whether a deal is an asset sale, the relative size of the target to the bidder, and whether the target is in a different industry than is the acquirer. To further control for asymmetric information about the value of the target, we also include in the Table VI, Panel A models the indicator variables for if the target operates in an industry with high earnings volatility or with low analyst coverage. To partly control for the issue that the sellers of unlisted targets could have urgent liquidity needs and that these needs can affect the discount in the price paid for an unlisted target (e.g., Officer (2007)), we include as an independent variable whether the method of payment consists only of cash.Also, because Officer (2007) shows that discounts for unlisted targets are larger when credit conditions are tighter, we control for the average spread of commercial and industrial loan rates over the federal funds rate during the four quarters prior to the quarter of the deal completion date. Finally, as in our prior multivariate analyses, we control for deal value and industry and year fixed effects.

The coefficient on the escrow agreement indicator variable in first model in Table VI, Panel A is significant and positive. We interpret the coefficient of 0.084 to suggest that, on average, the discount in the deal-value-to-sales multiple for an unlisted target relative to the median deal-value-to-sales multiple of comparable public targets is reduced by 8.4% through the use of an escrow contract. For instance, for the typical unlisted target in our sample that has a discount in its deal-value-to-sales multiple of 24.3%, the use of an escrow contract would reduce this discount to 15.9%. The economic impact of using an escrow contract on the discount in the price paid for an unlisted target is slightly higher if instead we use the regression coefficient on the instrumented escrow agreement indicator variable in the second model of Table VI, Panel A, (i.e., the coefficient of 0.202). We multiply this coefficient with the mean value for the instrumented escrow indicator variable of 0.544 and estimate that this coefficient suggests a reversal of the unlisted target price discount of 11.0% if an escrow contract is included in the acquisition agreement. We do not include a control variable in the Table VI, Panel A models for whether the seller of an unlisted target is financially constrained that is constructed using data about the seller because this data is not available for acquisitions of stand-alone private firms. Also, as previously noted, in the majority of the deals for subsidiary targets the parent firm of the subsidiary is a privately held firm and would consequently not be included in the Compustat database.

21 of 26

21

Economic Effects

OLS estimate shows reduction by 8.4% in discount (24.3%) to 15.9%.

2SLS estimate shows reduction by 11.0% in discount (24.3%) to 13.3%.

Such estimates do not account for claims from escrow fund by bidder after closing.

- Escrow Services at J.P. Morgan Chase report average 60% of escrow

funds returned to seller.

- This implies that 4.9% (=0.4*12.2%) of sales proceeds are kept by bidder.

- Reversal in discount is approximately:

3.5% = 8.4% -4.9% (OLS)

6.1% =11.0% - 4.9% (2SLS)

The coefficient on the escrow agreement indicator variable in first model in Table VI, Panel A is significant and positive. We interpret the coefficient of 0.084 to suggest that, on average, the discount in the deal-value-to-sales multiple for an unlisted target relative to the median deal-value-to-sales multiple of comparable public targets is reduced by 8.4% through the use of an escrow contract. For instance, for the typical unlisted target in our sample that has a discount in its deal-value-to-sales multiple of 24.3%, the use of an escrow contract would reduce this discount to 15.9%. The economic impact of using an escrow contract on the discount in the price paid for an unlisted target is slightly higher if instead we use the regression coefficient on the instrumented escrow agreement indicator variable in the second model of Table VI, Panel A, (i.e., the coefficient of 0.202). We multiply this coefficient with the mean value for the instrumented escrow indicator variable of 0.544 and estimate that this coefficient suggests a reversal of the unlisted target price discount of 11.0% if an escrow contract is included in the acquisition agreement.

Because when an escrow contract is used the bidder might ultimately keep some of the escrow account funds, we also provide an economic interpretation of the reversal of the unlisted target discount caused by the use of an escrow contract after accounting for this possibility. From bidders’ public disclosures it is not possible to determine the fraction of the escrow account funds that are returned to the seller parties. However, the escrow services division of J.P. Morgan Chase reported to us that in the context of escrow contracts used in acquisitions of unlisted targets, on average, 60% of escrow account funds are returned to the seller parties. In Panel A of Table I, we document that for the 491 deals in our sample that include an escrow contract that, on average, 12.2% of the sale proceeds are deposited into an escrow account. Thus, based on the assumption that typically 60% of the funds in the escrow account are returned to the seller parties, we estimate that, on average, among the deals in our sample that use an escrow contract 4.9% (=0.4*12.2%) of the sale proceeds are ultimately kept by the acquiring firm. Given that the coefficient estimates on the escrow indicator variable and the instrumented escrow indicator variable in Table VI, Panel A suggest that the use of an escrow contract leads to a reversal of the unlisted target discount of approximately 8.4% or 11.0%, this implies that after accounting for funds in the escrow account that are not returned to the sellers that the use of an escrow contract leads to a reversal in the unlisted target discount of approximately 3.5% to 6.1% (=8.4% - 4.9% and 11.0% - 4.9%). Overall, the results for the first two models in Panel A of Table VI are consistent with the use of an escrow contract reducing a bidder’s due diligence costs and with the bidder passing on to the sellers some of the created real value in the form of a higher final price for the unlisted target.

In the third to sixth models in Panel A of Table VI we examine separately the effect of using an escrow contract on the discount in the price multiple for a private firm or subsidiary target relative to the price multiple of comparable publicly traded targets. In the third and fifth models we regress the discount in the price multiple on the indicator variable for whether an escrow contract is used. In the fourth and sixth models we replace the indicator variable for whether an escrow contract is used with fitted values for whether an escrow contract is used that are calculated from the second and third models in Table IV, that separately examine private firm and subsidiary targets.

22 of 26

22

Economic Effects

Estimates higher for stand-alone targets:

- discount 17.1%.

- reversal 12.6%-14.3% (before deducting funds kept by bidder).

- net reversal is 7.4%-9.1% (net of 0.4*13% funds kept by bidder).

Estimates lower for subsidiary targets:

- discount 33.3%.

- reversal 4.4%-5.8% (before deducting funds kept by bidder).

- net reversal is 0.8%-2.2% (net of 0.4*8.9% funds kept by bidder).

On board:

14.3% = 0.228 * 0.626 (for private stand alone)

5.8% = 0.139 * 0.415 (for subsidiary targets)

The actual estimates for the effect of the escrow on stand-alone and subsidiary are as follows:

For stand-alone:

12.6% - 0.4*13% = 7.4%

14.3%- 0.4*13% = 9.1%

For subsidiary:

4.4% – 0.4*8.9% = 0.84%

5.8% - 0.4*8.9% = 2.24%

0.4*13% = 5.2% (the returned funds to bidder in stand-alone deal)

0.4*8.9% = 3.56% (the returned funds to bidder in subsidiary deal)

The findings for the third to sixth models show that escrow contracts reduce the discount in the price paid for an unlisted target both in the case of private firm and subsidiary targets. However, the effect is substantially larger for private firm than for subsidiary targets. The results for the third and fifth models suggest that the discount in the price paid for private firm and subsidiary targets relative to that for comparable public targets is reduced, respectively, by 12.6% and 4.4% through the use of an escrow contract. In the fourth model the coefficient estimate on the instrumented escrow agreement indicator is 0.228. We multiply this coefficient with the mean value for the instrumented escrow indicator variable for private firm targets of 0.626 (calculated from the second model in Table IV) and estimate that this coefficient suggests a reversal of the unlisted target discount of 14.3% for a private firm target if an escrow contract is included in the acquisition agreement. Similarly, we multiply the coefficient estimate on the instrumented escrow agreement indicator in the sixth model of 0.139 with the mean value for the instrumented escrow indicator variable for subsidiary targets of 0.415 (calculated from third model in Table IV) and estimate that this coefficient suggests a reversal of the unlisted target discount of 5.8% for a subsidiary target if an escrow contract is included in the acquisition agreement.

23 of 26

23

Analysis of Bidder Announcement Returns (-1 days to +1 days)

Table VII. Panel A.

	All Targets		Stand-alone Targets		Subsidiary Targets
	OLS	2SLS	OLS	2SLS	OLS	2SLS
Escrow Agreement Indicator	0.014^*		0.016^**		0.014^*
(t-stat)	(1.92)		(1.96)		(1.80)
Instrumented Escrow Agreement Indicator		0.033^*		0.037^**		0.019
		(1.85)		(1.94)		(1.57)
Target is a Stand-alone Private Firm	0.002	-0.008
	(0.26)	(0.83)
Acquisition is Considered an Asset Sale	0.006	0.009	-0.015	-0.014	0.023^**	0.034^**
	(0.79)	(1.09)	(1.34)	(1.01)	(2.13)	(2.81)
Relative Size of Target to Acquirer	0.012	0.020^**	0.022^*	0.024^**	0.005	0.020
	(1.55)	(2.04)	(1.77)	(1.97)	(0.44)	(1.18)
Target & Acquirer in Different Industries	-0.006^*	-0.009^**	-0.010^**	-0.012^**	-0.001	-0.004
	(1.79)	(2.24)	(2.03)	(2.15)	(1.05)	(1.36)
Log(Market Value of Acquirer Assets)	-0.004	-0.006	-0.005	-0.007	-0.002	-0.001
	(1.20)	(1.41)	(0.91)	(1.19)	(0.42)	(0.26)
Top Quintile Earnings Volatility Target Industry	-0.015	-0.017	-0.026^*	-0.025^*	-0.006	-0.009
	(1.28)	(1.34)	(1.71)	(1.83)	(0.36)	(0.46)
Low Analyst Coverage Target Industry	-0.005	-0.001	-0.014	-0.001	-0.004	-0.021
	(0.48)	(0.05)	(0.96)	(0.08)	(0.21)	(0.93)
Method of Payment is Only Cash	0.006	0.001	0.009	0.011	0.002	0.001
	(0.88)	(0.08)	(0.95)	(0.09)	(0.15)	(0.07)
Method of Payment is Only Stock	0.038^**	0.041^**	0.046^**	0.049^**	-0.006	-0.020
	(2.29)	(2.33)	(2.38)	(2.51)	(0.01)	(0.57)
Commercial & Industrial Loan Rate Spread	0.009	0.015	0.001	0.012	0.015	0.026
	(1.08)	(1.48)	(0.02)	(0.85)	(1.01)	(1.29)

Observations	844	783	504	483	340	300
Adjusted R-squared	0.019	0.011	0.047	0.040	-0.013	-0.014

Fama-French 49 industry (of the acquirer) fixed effects and year fixed effects are included in all the models.

We investigate whether the use of an escrow contract in the acquisition of an unlisted target leads to more positive market reactions for bidders at the announcement of an acquisition deal. We expect that this should be the case if not all of the cost savings for a bidder from using an escrow contract are passed on to the sellers of the target in the form of a higher acquisition price.

Table VII, Panel A provides the results of an examination of this issue. Bidder announcement returns are the (-1, 1) day abnormal return surrounding the acquisition announcement date, calculated using the market model where the estimation window is the (-220, -20) days prior to the acquisition announcement date. The acquisition announcement date is the date of the filing of form DEF 14A for the preliminary acquisition deal. In this filing, it would be reported if an escrow contract is included in the acquisition agreement given that the escrow contract would be noted as part of the price consideration offered in the transaction.

Although not tabulated, we find that mean (median) bidder acquisition announcement returns for the full sample and for the subsamples consisting only of private firm targets or of subsidiary targets are, respectively, 3.0% (2.1%), 2.9% (2.2%), and 3.2% (2.0%). The magnitude of these announcement returns is similar to that documented in prior work that has examined bidder announcement returns in acquisitions of private firm and subsidiary targets (e.g., Fuller, Netter, and Stegemoller (2002)).

In the Table VII, Panel A models the dependent variable is the bidder acquisition announcement return. The independent variables in these models are the same as in Table VI, except that because the dependent variable relates to the bidder we control for bidder size rather than deal size. Also, we additionally control for whether the method of payment is only stock because prior work shows that in acquisitions of unlisted targets bidder announcement returns tend to be more positive when the method of payment is only stock (e.g., Chang (1998), Fuller, Netter, and Stegemoller (2002), and Officer, Poulsen, and Stegemoller (2009)).

The results for the first two models in Panel A of Table VII show that the use of an escrow contract is associated with higher bidder acquisition announcement returns and that this is the case both using OLS and a 2SLS approach that controls for the potential endogeneity of having an escrow contract in place. These findings are consistent with escrow contracts creating value for bidders by reducing their due diligence costs in private firm and subsidiary acquisitions.

In the third and fourth models of Panel A of Table VII we report the results for only private firm targets, while in the fifth and sixth models we report the results for only subsidiary targets. The results for the OLS estimations in the third and fifth models show that irrespective of whether a target is private firm or subsidiary target that the use of an escrow contract is associated with higher bidder acquisition announcement returns. The findings for the 2SLS estimates in the fourth model document that for acquisitions of private firm targets when we account for the potential endogeneity of having an escrow contract in place that the association between the use of an escrow contract and bidder acquisition announcement returns continues to be positive. However, for subsidiary targets we find that after accounting for this potential endogeneity that although there remains a positive association between the use of an escrow contract and bidder announcement returns this association becomes insignificant at conventional significance levels (t-statistic = 1.57).

The adjusted R-squared values for the fifth and sixth models in Table VII are negative. We find that this is due to the large number of control variables with insignificant coefficients in these two models. Specifically, if we rerun the two models including the escrow agreement indicator variables and the only control variable with a significant coefficient (i.e., the indicator for whether the transaction is an asset sale) and year and industry fixed effects the adjusted R-squared values increase to 0.005 and 0.013, and the escrow agreement and asset sale indicator variables retain their significant coefficients. However, we note that in these two models we present the results using the extended list of controls for consistency with our prior analyses.

24 of 26

24

Implications

If both bidders & sellers gain on average by using escrow contracts, shouldn’t all unlisted target acquisitions use it?

Potential reasons for why these contracts are not used in all deals:

Sellers who are aware of potential misstatements in reps. and warranties.

Discord among target shareholders with respect to including an escrow.

Sellers with urgent need of liquidity (escrows reduce liquidity).

Positive association b/n use of escrow contract & bidder acquisition announcement returns does not imply all bidders should use escrow contract.

Our evidence suggests that because the use of an escrow contract reduces a bidder’s due diligence costs, the use of such a contract enables sellers of unlisted targets to obtain a higher sales price and positively impacts the extent to which acquisitions of such targets leads to value creation for the bidder. A potential question that may arise is why these contracts are not used in all deals involving unlisted targets. Below we discuss several potential reasons for why these contracts are not used in all such acquisition deals.

In some cases sellers who are aware of potential misstatements made to the bidder in the representations and warranties section of the acquisition agreement may prefer to not include an escrow contract within the acquisition agreement. If so, then sellers’ preference to not include an escrow contract in the acquisition agreement could, at times, result in bidders having larger concerns about the overall quality of the target and to a larger discount in the price paid for the unlisted target. This scenario would be similar to that discussed in prior work (e.g., Akerlof (1970) and Myers and Majluf (1984)), in which buyers will demand a discount in the price paid for a product if it is difficult to verify the product’s quality and they are not offered a warranty for it. Also, under this scenario because the bidder would not be able to reduce its due diligence costs through the use of an escrow contract, this could result in lower bidder acquisition announcement returns.

Another reason why an escrow contract might not be used in the acquisition of an unlisted target is that there could be discord among target shareholders with respect to including an escrow contract in the acquisition agreement. For instance, in some cases a number of smaller target shareholders, whose combined ownership is substantial, could prefer to not include an escrow contract in the acquisition agreement and instead make one or two large shareholders of the target more liable for seller breaches of warranties and representations made to the bidder. Finally, another factor that could result in an escrow contract not being used in the context of a given private firm or subsidiary acquisition is that because at times an important motive for these sales is that the sellers have an urgent need for liquidity (e.g., Officer (2007)), in some cases it could be important for the sellers to have immediate access to the entire sale proceeds. In these instances, an escrow contract could be viewed by the sellers as costly, as it would reduce the liquidity that they seek. It is important to point out that the result of a positive association between the use of an escrow contract and bidder acquisition announcement returns does not imply that all bidders should attempt to use an escrow contract when acquiring an unlisted target. Instead, this result merely captures the issue that, on average, those bidders who do use an escrow contract do so efficiently to raise firm value.

We note that this scenario is not inconsistent with the Table 4 finding that, on average, the existence of a dominant target shareholder is positively associated with whether an escrow contract is included in the acquisition agreement.

25 of 26

25

Conclusions

Escrow contracts are commonly used in unlisted target acquisitions.

Used in 52.1% of all unlisted target acquisitions.
Common for stand-alone targets (75.6% of escrows) and stock purchases (80.0% of escrows).

Efficient contracting mechanism, aimed at reducing transaction risk and reducing informational asymmetry and related due diligence costs.

Has important valuation impact on target & increases bidder acquisition announcement returns.

On average, reduces unlisted target discount by 3.5-6.1%.

Overall, our paper makes several contributions.

First, we document that escrow contracts are frequently used in acquisitions of private stand-alone firm and subsidiary targets and the results of our tests on the determinants and effects of these contracts shed light on how the participants in these transactions solve contracting problems. This is important because although much is known about acquisitions of publicly traded targets, our understanding of the process in private firm and subsidiary acquisitions is relatively limited.

Second, our findings provide new insights on the determinants of the method of payment used in corporate acquisitions and how it can be a function of buyers and sellers trying to manage acquisition-related transaction risk.

Finally, our paper also contributes to the broader literature on how financial contracting can create real value and resolve problems resulting from information asymmetry and moral hazard (e.g., Myers and Majluf, (1984), Grossman and Hart (1986), Hart and Moore (1990), Kaplan and Stromberg (2003)).

Related work shows that the use of stock to pay for an acquisition can reduce bidder acquisition-related transaction risk resulting from information asymmetry about the value of the target (Hansen (1987) and Officer, Poulsen, and Stegemoller (2009)) and that the use of collars in acquisitions that are at least partly paid for with stock lower sellers’ transaction risk resulting from volatility in a bidder’s stock price (Houston and Ryngaert (1997) and Officer (2004)). As well, prior work reports that the use of earnout payments helps to mitigate bidder transaction risk by providing incentives to a target firm’s managers to remain with the firm post-acquisition and maximize acquisition related synergies (Kohers and Ang (2000), Datar, Frankel, and Wolfson (2001), and Cain, Denis and Denis (2011)). Prior work documents that the method of payment is also determined by factors such as taxes (e.g., Kaplan (1989), Brown and Ryngaert (1991), and Erickson (1998)), a firm’s ownership structure (e.g., Amihud, Lev, and Travlos (1990), Martin (1996), and Ghosh and Ruland (1998)), firm financing policies (e.g., Bharadwaj and Shivdasani (2003) and Harford, Klasa, and Walcott (2009)), and bidder firm overvaluation (e.g., Shleifer and Vishny (2003), Rhodes-Kropf, Robinson, and Viswanathan (2005), and Dong, Hirshleifer, Teoh, and Richardson (2006)).

26 of 26

26

Claims Post-closing (source: J.P. Morgan Chase Escrow Services, 2009)

(Show if time permits)