Gordon S. Roberts

Designing an immunized portfolio: Is M-squared the key?

Abstract The theoretical and empirical properties of M -squared, a measure of cash flow dispersion used in designing duration-hedged portfolios, are examined. Contrary to prior research, minimizing M -squared is not independent of the stochastic process and the minimum M -squared portfolio is a ‘bullet’ only under a specific, convexity condition derived in the paper. Using a data base of default-free, Government of Canada bonds to set up minimum M -squared, duration-matching portfolios, we find that the convexity property does not hold in general and that minimum M -squared portfolios fail to hedge as effectively as portfolios including a bond maturing on the horizon date.

A note on market response to corporate loan announcements in Canada

Abstract This note validates the key results of prior studies of bank loan announcement effects using a common data set drawn from the Canadian capital market. Announcements of bank loans are associated with positive abnormal returns significantly higher than for private placements and loan syndications. Announcement effects are most pronounced when monitoring is most intense and when an announcement signals that the banks private information is favorable. Conclusions of prior studies on bank loan announcements, conducted exclusively on US data, are robust for a different banking system.

Duration for bonds with default risk

Abstract Does approximating duration estimates by ignoring default risk lead to error in the two major duration applications — measuring interest — rate price elasticity and immunization? We derive a general expression for duration in the presence of default risk based on Jonkharts term structure model (On the term structure of interest rates and the risk of default, Journal of Banking and Finance 3, 253–262, 1979) extended to encompass risk aversion. The model includes terms for default probabilities and default payoffs in each period as well as for a delay between the occurrence of default and the final default payoff. Our main conclusion is that practical duration applications involving bonds with default risk must employ duration measures adjusted for default risk.

Bond portfolio immunization: Canadian tests

Abstract Based on findings derived from a data base of default-and option-free Government of Canada bonds, our main conclusion is that duration-matching strategies succed in reducing interest rate risk more effectively than strategies that call for matching bond maturity to the length of the planning horizon. The present article stands alone in immunization research in establishing an advantage to duration matching even in the absence of term structure fitting. For this result to stand, care must be taken in the design of appropriate duration-matching strategies.

Default- and call-adjusted duration for corporate bonds

Abstract Call and default can potentially alter the timing and amounts of promised cashflows for callable, corporate bonds. While prior research has indicated the theoretical importance of adjusting Macaulay duration for the impacts of default and call, the question of their relative impact remains a matter of debate [The High Yield Debt Market, Dow Jones Irwin, New York, 1990, p. 18; J. Finan. 53 (1998) 2225]. We develop a theoretical analysis incorporating both default and call effects on duration and test its implications employing a previously unexplored data base of Canadian, investment grade, corporate bond indices containing an unusual provision making it possible to identify callable and noncallable indices.

Captive Finance Subsidiaries and the M-Form Hypothesis

Williamsons theory of hierarchy is used to show that the widespread and, before now, unexplained corporate practice of establishing captive finance subsidiaries may be understood as an instance of multidivisonal-form reorganization. It is argued that such reorganization enhances internal efficiency and eases the monitoring task of lenders, thus facilitating borrowing. The debt ratios of firms in five U.S. industries are examined and the evidence, while not conclusive, provides support for the hypothesis. Thinking of captive finance subsidiaries in terms of the theory of hierarchy is consistent with the rationale for finance subsidiaries presented by corporate officers.

Loan Rates and Collateral

We investigate the relation between corporate loan spreads and collateralization. We use propensity scoring to create a matched sample of pairs of loan facilities from the Dealscan database. We find that noncollateralized loans are associated with lower spreads even after controlling for risk.

Does the Secondary Loan Market Reduce Borrowing Costs

We show that lenders make price concessions for the right to resell loans and reveal a strong countervailing association between the ex ante probability of loan resale and the initial loan spreads. We disentangle the side effects (reduced monitoring) from the benefits (enhanced liquidity) brought by the secondary loan resales. The average net impact of simultaneously reducing the probability of the presence of resale constraint and raising the probability of resale across the full sample is to lower spreads by 14 basis points. On balance, the secondary loan market provides clear benefits to the issuers of debt.

Perspectives on Canadian Bank Insolvency during the 1930s

Jack Carr, Frank Mathewson and Neil Quigley (1995) (CMQ) introduce new archival evidence to challenge the hypothesis that Canadian banks enjoyed considerable capital forbearance during the 1930s (Lawrence Kryzanowski and Gordon S. Roberts, 1993) (KR). This note examines what the CMQ evidence has to tell us once opportunity-cost valuation and agency costs in government and the accounting profession are stirred vigorously into the inferential stew. The authors show that none of the CMQ findings is inconsistent with the original KR argument on capital forbearance. More broadly, they demonstrate how serious economic error can result from taking accounting and contractual formalisms at their face value.

Beta Instability When Interest Rate Levels Change

Boquist, Racette, and Schlarbaum [3] and Livingston [6] show that a security systematic risk may be expressed as a function of its duration. These results have led to research examining the role of duration in explaining systematic risk, but Lanstein and Sharpe [5] indicate that Livingstons expression relies on the implicit assumption that extra-market covariances between securities are insignificant. Lanstein and Sharpe argue that such an assumption is unwarranted. They find a significant negative relationship between extra-market covariances and differences in duration between paired samples of common stock. Their paper suggests that duration may be associated with unsystematic risk and that any relation between duration and systematic risk is more complex than implied in [3] and [6].