Financing Through Asset Sales

The following post comes to us from Alex Edmans and William Mann, both of the Department of Finance at the University of Pennsylvania.

In our paper, Financing Through Asset Sales, which was recently made publicly available on SSRN, we analyze a source of financing that is first-order in reality but relatively unexplored in the literature — selling non-core assets such as a division or a plant. Asset sales are substantial in practice: in 2010, there were $133bn of asset sales in the U.S., versus $130bn in seasoned equity issuance. In contrast, most existing research on a firm’s financing decisions studies the choice between debt and equity and ignores asset sales. We build a model that allows asset sales to be undertaken not only to raise capital, but also for operational reasons (dissynergies). We study the conditions under which asset sales are preferable to equity issuance and vice-versa, how financing and operational motives interact, and how firm boundaries are affected by financial constraints.

The firm comprises a core asset and a non-core asset. The firm must raise financing to meet a liquidity need, and can sell either equity or part of the non-core asset. Following Myers and Majluf (1984) (MM), we model information asymmetry as the principal driver of this choice. The firm’s type is privately known to its manager and comprises two dimensions. The first is quality, which determines the assets’ standalone (common) values. The value of the core asset is higher for high-quality firms. The value of the non-core asset depends on how we specify the correlation between the core and non-core assets. With a positive (negative) correlation, the value of the non-core asset is higher (lower) for high-quality firms. The second dimension is synergy — the additional value that the non-core asset is worth to its current owner.

It may seem that asset sales can already be analyzed by applying the general principles of MM’s security issuance model to assets, removing the need for a new theory specific to asset sales. Such an extension would suggest that assets are preferred to equity if they exhibit less information asymmetry. While information asymmetry is indeed an important consideration, our model identifies three new forces that also drive the financing choice and may outweigh information asymmetry considerations.

First, an advantage of equity is that new shareholders obtain a stake in the entire firm. This includes not just the core and non-core assets in place (whose value is unknown), but also the cash raised (whose value is known). This mitigates the information asymmetry of the assets in place — the certainty effect. In contrast, an asset purchaser does not share in the cash raised, and thus bears the full information asymmetry associated with the asset’s value. Hence, contrary to MM, even if equity exhibits more information asymmetry than the non-core asset, the manager may sell equity if enough cash is raised that the certainty effect dominates. Contrary to conventional wisdom, equity is not always the riskiest claim: if a large amount of financing is raised, equity becomes relatively safe.

Formally, a pooling equilibrium is sustainable where all firms sell assets (equity) if the financing need is sufficiently low (high). The choice of financing thus depends on the amount required. This dependence contrasts standard financing models, where the choice depends only on the inherent characteristics of the claim being issued (such as its information asymmetry (MM) or misvaluation (Baker and Wurgler (2002)) and not the amount required — unless one assumes exogenous limits such as notions of debt capacity. This result also has implications for the investment literature, in which disinvestment occurs due to financial constraints and so a greater financing need leads to more assets being sold. We show that a greater financial shock may reduce asset sales, as firms substitute into equity. Thus, such a shock can improve real efficiency, as firms hold onto synergistic assets and instead sell equity.

The certainty effect applies to any use of cash whose expected value is uncorrelated with firm quality: retaining it on the balance sheet to replenish capital, repaying debt, paying dividends, or financing an uncertain investment whose expected payoff is independent of firm quality. We also analyze the case in which the investment return is correlated with firm quality, and thus exhibits information asymmetry. It may appear that the certainty effect should weaken, since the funds raised are no longer held as certain cash. This intuition turns out to be incomplete, because there is a second effect. Since investment is positive-NPV, it increases the value of the capital that investors are injecting. If the desirability of investment (for both low- and high-quality firms) is high compared to the additional return generated by the high-quality firm over the low-quality firm, the second effect dominates. Somewhat surprisingly, the certainty effect can strengthen when cash is used to finance an uncertain investment. This effect makes equity easier to sustain. In contrast, if the additional return generated by the high-quality firm is sufficiently large, then asset sales become preferable. Due to the role of the investment return, the source of financing depends on the use of financing. In almost all cases, it remains robust that asset (equity) sales are used for low (high) financing needs.

A second driver of the financing decision is the level of synergies between the non-core asset and the firm. This consideration leads to “threshold” semi-separating equilibria, in which a firm sells assets if synergies are below a cutoff and equity otherwise. While models of firm boundaries also predict that firms will sell dissynergistic assets, here these operational motives interact with financing/market timing reasons. Some high-quality firms sell assets not because they are low quality, but because they are dissynergistic, allowing low-quality firms to pool with them. They can disguise an asset sale driven by overvaluation (the asset is of low quality and has a low common value) as instead being driven by operational reasons (it is dissynergistic and only has a low private value) — the camouflage effect. Low-quality asset sellers can make greater profits than in the pooling equilibria, where the financing choice does not depend on synergies and so no disguise is possible. A market in which firms are selling assets for operational reasons is “deep” and allows other firms to exploit their private information by selling overvalued assets. This notion of “market depth” is similar to the Kyle (1985) model of securities trading, where a deep market arises when liquidity traders are selling their securities for reasons other than a low common value. Such depth allows informed traders to profit from selling securities that do have a low common value.

A third driver is the correlation effect, which represents an advantage to selling assets. When a firm issues equity, it suffers an Akerlof (1970) “lemons” discount — the market infers that the equity is low-quality from the firm’s decision to issue it. Not only does the market pay a low price for the equity issued, but also it attaches a low valuation to the rest of the firm, because it is perfectly correlated with the issued equity. When a firm sells non-core assets, it also receives a low price, but critically this need not imply a low valuation for the firm as the asset sold may not be a carbon copy. Thus, firms can sell poorly performing assets without sending a negative signal. Formally, in the negative correlation model, the parameter values that support the equity-pooling equilibrium are a strict subset of those that support the asset-pooling equilibrium. An implication is that conglomerates issue equity less often than firms with closely related divisions. In addition, asset sales (equity issuance) should lead to positive market reactions, as found empirically. The analysis also highlights a new benefit of diversification: a non-core asset is a form of financial slack. While the literature on investment reversibility (e.g. Abel and Eberly (1996)) models reversibility as an inherent feature of the asset’s technology, here an investment that is not a carbon copy of the firm is “reversible” in that it can be sold without negative inferences.

The full paper is available for download here.

Both comments and trackbacks are currently closed.