1 / 27

Internal Capital Markets

Internal Capital Markets. In the 1990s researchers began investigating resource allocation decisions within the firm using Compustat data and other sources

albert
Download Presentation

Internal Capital Markets

An Image/Link below is provided (as is) to download presentation Download Policy: Content on the Website is provided to you AS IS for your information and personal use and may not be sold / licensed / shared on other websites without getting consent from its author. Content is provided to you AS IS for your information and personal use only. Download presentation by click this link. While downloading, if for some reason you are not able to download a presentation, the publisher may have deleted the file from their server. During download, if you can't get a presentation, the file might be deleted by the publisher.

E N D

Presentation Transcript

  1. Internal Capital Markets • In the 1990s researchers began investigating resource allocation decisions within the firm using Compustat data and other sources • Internal capital markets are a major channel of capital allocation in modern industrial economies; managers in every firm must allocate capital across different projects • Lamont (1997) notes that between 1981 and 1991, internal funds account for more than 75% of capital outlays for U.S. nonfinancial corporations • Related issues include: • The choice of whether to rely on internal or external finance to generate funds for investments • Whether diversification within the firm is a value maximizing strategy or due to agency problems • The extent to which internal capital markets allocate resources efficiently across divisions

  2. The Evolution of Research on Internal Capital Markets • Internal capital markets may differ from external markets due to differences in information, incentives, asset specificity, control rights, or transaction costs • We’ll cover most of the main contributions to this area in rough chronological order • Because the research has developed fairly recently, internal capital markets is a good case study for examining how research agendas evolve over time • We’ll begin with three key papers: Gertner, Scharfstein, and Stein (1994), Lamont (1997), Stein (1997)

  3. Gertner, Scharfstein, and Stein (1994) • This paper focuses on the costs and benefits of external vs. internal financing in theory • The authors focus primarily on comparing an internal capital market with bank lending • They argue that the internal capital allocation process has three important consequences: • It leads to more monitoring than bank lending • It reduces managers’ entrepreneurial incentives • It makes it easier to efficiently redeploy the assets of projects that are performing poorly under existing management

  4. Theory • Gertner, Scharfstein, and Stein (1994) build on the general framework of Grossman and Hart (1986), in which ownership is associated with residual control rights over the use of firm’s assets • Thus, their key assumption is that in an internal capital market the residual control rights reside with the capital supplier (corporate headquarters), while in a bank lending arrangement they reside with the firm’s manager (not the bank) • Residual control rights affect the incentives and ability of the capital provider and the incentives of the division managers • In a world with complete contracts there would be no difference between internal and external financing

  5. Results • The capital provider has more incentive to monitor when it has residual control rights because the rights give it more of the gains from monitoring (better flow of information between users and providers of capital) • Giving control rights to capital providers reduces the division manager’s incentives because the manager is more vulnerable to opportunistic behavior by headquarters. The manager may not get all of the rents from his efforts, which reduces his incentives • When headquarters owns multiple related business units, it can redeploy assets across units in response to profit opportunities. In contrast, an external provider would have to sell the assets to another user to accomplish this objective and may not receive the full value of the assets

  6. Possible Extensions • The authors’ limited focus on residual control rights prevents them from analyzing agency problems at the level of the capital provider • For example, is headquarters more inclined to make inefficient investments than a bank lender? • One reason why this might occur follows from the headquarter’s extra monitoring and residual control rights • Monitoring involves more interaction with the division managers; this increases the scope for influence activities where the managers try to distort resource allocation in their favor • Another issue is do liquidity problems in one division spill over to the firm’s other divisions? This is the question that Lamont (1997) investigates empirically

  7. Lamont (1997) • This paper uses data from the 1986 oil price decrease to examine the capital expenditures of non-oil subsidiaries of oil companies • He tests the joint hypotheses that • A decrease in cash/collateral decreases investment, holding fixed the profitability of investment • The finance costs of different parts of the same corporation are interdependent • The results support the joint hypothesis: oil companies significantly reduced their non-oil investment compared to the median industry investment • The 1986 decline in investment was concentrated in non-oil units that were subsidized by the rest of the company in 1985

  8. Empirical Approach • The author uses the Compustat database to identify a group of firms that have corporate segments in the oil extraction industry and in non-oil industries, where “non-oil” is defined as an industry with profits that are not positively correlated with the price of oil • Then he tests the hypothesis: do large cash flow/collateral value decreases in a firm’s oil segments decrease investment in the non-oil segments? Standard finance theory suggests they should not • He focuses on the 1986 oil shock when oil prices fell by 50% because this event seems unambiguously exogenous to any firm • His test can be framed as a joint hypothesis: • Did the oil shock affect the cost of finance for oil segments? • Did the effect spill over to affect the cost of finance in non-oil segments?

  9. Implications • A simple perfect capital markets model implies that when a company’s oil segment cash flow falls, the same company’s non-oil segment should be unaffected if the net present value of non-oil investment is unaffected • The empirical support for the joint hypothesis suggests that external capital markets are imperfect (so firms rely on internal capital markets) and that different parts of a firm are interdependent • The findings are consistent with previous research that suggests that diversified companies tend to subsidize and over invest in poorly performing segments

  10. Empirical Issues • A large literature documents the empirical relationship between liquidity and investment; there is a strong correlation between a firm’s cash and its investment • However, causation is not clear; both investment and cash flow are driven by underlying shocks to profitability • Existing studies have attempted to control for the profitability of investment by including a measure of Tobin’s q • Since exogenous instruments for cash that are uncorrelated with the profitability of investment are hard to find, researchers have focused on examining the differences in cash-investment correlations between groups of firms hypothesized to have different dependence on internal finance

  11. Previous Approaches Previous studies use panel data on firms to estimate: I/K = a + b Q + c Cash/K + year dummy + firm dummy + error I is investment K is the capital stock at the beginning of the period Q is Tobin’s q CASH is a measure of cash flow or cash stock To test the hypothesis that two groups of firms face different finance constraints, the coefficient c is compared across different groups (or perhaps different time periods with different credit conditions) This comparison is imperfect because firms might differ within groups and q may be a poor proxy for investment opportunities

  12. Lamont’s Approach • Lamont attempts to find an exogenous instrument for cash through a natural experiment • A key innovation is to use corporate segment level data • Publicly owned firms in the U.S. are required to report certain data disaggregated by segment, with a segment for each industry the company participates in • The null hypothesis is that corporate segments operate as stand-alone units, there is no role for an internal capital market, and each segment finances its investment from its own internal finance or from external finance secured by its own collateral • In this view, corporations operate multiple lines of business solely to exploit product market synergies or because their managerial talent can add value across a range of activities

  13. The Oil Shock of 1986 • In late 1985 Saudi Arabia increased its oil production and within five months crude oil prices had fallen by 50% • Profit rates for the oil and gas production/extraction industries fell dramatically

  14. Data • Accounting standards require corporations to report five annual variables at the segment level for each segment that constitutes at least 10% of total sales and which are in a different industry than the rest of the corporation: sales, operating profit, capital expenditures, depreciation, and identifiable total assets • Compustat reports these five items along with SIC codes for each segment • This data is far from perfect; firms might divide overhead costs and assets that provide benefits to multiple segments in an arbitrary way • Lamont focuses on normalized changes in segment-level capital expenditure, so any firm-specific practice that is constant over time should not affect the results • He examines every firm that in 1985 had at least 25% of their cash flow from the oil and gas extraction industry and focuses on those with substantial non-oil activities

  15. Sample • The final sample consists of 26 firms; most are large • There is a survivorship issue: • In the presence of imperfect capital markets, diversified firms would underinvest in their non-oil divisions • One way to mitigate this inefficiency and raise cash at the same time is to sell off these divisions; many firms did so in 1986 • Lamont’s sample contains only survivors; this may bias the results toward accepting the null

  16. Method • He does not observe physical capital K at the segment level • Given this, he focuses on the ratio of contemporaneous investment to contemporaneous sales: I/S • He focuses on change in I/S between 1985 and 1986 • He also cannot observe Tobin’s q (a proxy for investment opportunities) at the segment level • He uses industry-adjusted data to address this problem (he subtracts the industry median change in I/S including only non-oil firms from the firm’s change in I/S); the results are unaffected by whether he does the industry adjustment or not • The raw data shows that most non-oil segments experienced an increase in cash flow between 1985 and 1986 (in contrast to the oil industry) • However, capital expenditures declined in most segments

  17. Under- vs. Over-Investment • Principal agent models stress that diversified firms may overinvest; previous studies present evidence that supports this view • A reduction in free cash flow may discipline managers and reduce overinvestment; Lamont finds evidence that supports this view • In 1985, nonoil segments owned by oil companies under-performed their industry peers; profit to sales ratios were 2% lower than the industry median • However, I/S ratios were the same as the rest of the industry • In 1986, these nonoil segments performed about as well as the industry and I/S fell below industry levels • Thus, part of the reason why I/S fell may be that oil companies were investing too much in below average segments before the shock and then became more disciplined after the shock; regression analysis supports the view that oil performance mattered less for non-oil segments after the shock

  18. Extensions/Conclusions • In principle, Lamont’s approach could be used to measure the effect of any financial shock on diversified firms • Compustat data is not perfect; one problem is that the best proxy for cash flow is pretax operating income (not after-tax) • Lamont suggests obtaining plant-level data from the Census bureau • This data would add observations and supply more information about individual production units (plants) • He also suggests looking at longer time periods and asking whether negative and positive shocks have similar effects

  19. Stein (1997) • This is a theory paper that addresses three main questions: • Why use internal capital markets? • Why combine different projects in one firm? • What is the optimal size/scope of an internal capital market? • The main explanation in the paper is that headquarters has private information that outsiders do not have • Given this, HQ can pick winners inside the firm; one division’s assets can be used to raise funds and then the funds can be transferred to another division if needed

  20. Key Assumptions • There are two main implicit assumptions: • Adverse selection and moral hazard prevent external capital markets from working efficiently; HQ and project managers get utility from controlling resources, so they tend to overstate their prospects • HQ has the authority, desire, and ability to pick winners among divisions/projects

  21. The Basic Project-Level Agency Problem with External Financing • A founder has a patent but must obtain financing to develop a product • The founder also must hire a project manager; he cannot manage the project himself • Assume that the founder is a profit maximizer, that the rate of return required by financiers is 0, and that the reservation wage of the manager is 0 • The amount invested in the project can either be 1 or 2

  22. Returns • The return depends on the state of the world: In state B (bad): y1 if 1 was invested; y2 if 2 was invested Both y1 and y2 are verifiable, so they can be contracted on Assume 1 < y1 < y2 < 2, so if it is known that the state is B then it is optimal to invest 1 In state G (good): qy1 if 1 was invested; qy2 if 2, where q > 1 Assume qy2 – 2 > qy1 – 1, so if the state is G it is optimal to invest 2 • Ex ante, state G occurs with probability p • Assume that only the project manager observes the state

  23. Private Managerial Benefits • Managers want more resources • Assume that managers get private benefits that are unverifiable (so they cannot be contracted on): sy1, sy2 in state B; sqy1, sqy2 in state G, where s is a constant

  24. Credit Rationing • Suppose managers cannot credibly reveal their information about the state • Then financiers must either always invest 1 or always invest 2 • Note that by assumption, pqy1 + (1 – p)y1 >1, so it is always possible to raise 1 • Assume that financiers always invest 1 in this case (if p is small enough investing 1 is preferred to investing 2) • It follows that sometimes capital is inefficiently allocated (every time state G occurs)

  25. The Role of HQ Assumptions: • HQ can acquire information about projects’ ex ante prospects • HQ has no cash of its own • HQ has control rights which give it f of the private benefits of any project it oversees, and this surplus extraction dilutes the manager’s incentives (so all cash flows fall by a factor k) • HQ’s control rights give it the authority to redistribute resources across projects; this implies that HQ can raise funds for n projects and then divert those funds toward whichever projects it prefers (which might involve leaving some projects with no resources)

  26. The Role of HQ • Clearly HQ always reduces value if there is only one project – output is reduced by a factor k and there is no offsetting benefit • Consider the case where HQ has 2 projects, and assume that the project’s states are independent of each other and that HQ is no better at raising external funds than the managers would be (so it raises 2 units of capital) • Now the HQ can provide a benefit to the founder by picking winners: If project 1 is in state G and 2 is in state B, HQ can allocate both units of capital to project 1 Under q(y2 – y1) > y1 this is optimal; assume this condition holds • HQ has the incentive to reallocate capital because it derives benefits from both projects, whereas managers just care about their own project

  27. Diversification vs. Focus • Note that under the assumptions made so far, as the number of projects gets large, eventually proportion p are in state G and proportion (1 – p) are in state B; since the external financiers know p they can use this information to provide capital efficiently (provide 2 units of capital for proportion p of projects and 1 unit for the rest) • This result suggests that diversification is optimal; ideally HQ would have a large number of projects with independent shocks • Stein notes that if project evaluation errors can occur, then more focused firms may have an advantage if errors are correlated when projects are similar • The logic: if there are two projects and the true state is B,B, HQ is likely to observe either B,B or G,G; in either case it gives each project 1 unit of capital, which is optimal • HQ is unlikely to observe B,G or G,B because evaluation errors are correlated

More Related