The Theory Of Interest Kellison

Working With Kellison's Theory of Interest: What Actually Matters

The Theory of Interest by Stephen G. Kellison is the book most actuaries and financial mathematicians learned their interest theory from. It is not the only one, but it remains the standard reference for Exam FM and the foundation most people build everything else on. The book covers the basics — compound interest, annuities, bonds, duration, and immunization — and it does so in a way that is deliberately computational rather than abstract. That is its main characteristic and its main limitation. I spent years grading exam-level work and consulting on pricing models that used Kellison's frameworks. The thing nobody tells you about this book is that the notation takes up most of the cognitive load. Kellison writes every formula with subscripts and superscripts that look like alphabet soup until you've seen them twenty times. The actual math is straightforward algebra and geometric series, but the presentation forces you to translate between p-and-i notation and whatever shorthand your workplace uses. The method works like this. You start with the fundamental equivalence principle — money has a time value determined by an interest rate or discount factor. From there you build annuity formulas, then bond pricing, then measures of sensitivity like duration and convexity. The derivations are explicit. You are not expected to guess at a formula. Every result follows from the previous one through mechanical manipulation.

Here is where people run into trouble in practice. Kellison assumes exact payment timing and constant rates. Real cash flows are messy. I worked on a liability matching project where the obligations had variable dates depending on claim settlement patterns, and the textbook approach of treating everything as level annuities produced results that were off by several percentage points. The workaround was to decompose the liability stream into individual cash flow buckets and price each one separately using spot rates rather than a single yield. That took longer but it was the only way to get the immunization duration to line up within tolerance. Another common mistake is confusing nominal and effective rates without checking the compounding frequency. Kellison covers this in the first chapter, but people skip ahead. I have seen models fail because someone used an annually compounded rate as if it were a monthly one. It happens more often than you would think. The book's strengths are its worked examples and the problem sets. The examples are mostly clean and solvable. The problems range from routine to genuinely tricky, and working through them is what actually builds competence. Reading the chapters passively gives you the illusion of understanding. Doing the problems is where you find out what you do not know.

There are real downsides to relying solely on Kellison. The treatment of stochastic interest rates is thin. If you need to model rate movements for valuation or risk management, you will outgrow this book quickly. The section on duration and immunization is solid for static scenarios but does not address convexity adjustments in shifting environments the way more advanced texts do. For that you need something like McCallum or the later chapters of Shreve. Also, the book does not cover modern computational tools. Everything is designed to be done by hand or calculator, which is fine for exam preparation but not for actual production work where you would use a spreadsheet or Python script. If you are studying for Exam FM, Kellison is still the most efficient path. The coverage aligns closely with the syllabus, and the difficulty level matches the exam. Plan for roughly 150 to 200 hours of study if you are starting from scratch, with the heaviest weight on annuities, bonds, and duration. The early chapters move fast. Do not linger on them past the point of confidence. Most people waste three weeks on chapters one and two when they could be done in a few days. For practitioners, use Kellison as a reference rather than a primary text. Pull it when you need to verify a derivation or check a formula. Build your actual models in code. I keep a small library of Python functions that implement the core Kellison formulas — annuity present values, bond prices, Macaulay and modified duration, convexity — and call those instead of re-deriving anything. The lookup time is maybe thirty seconds per function once it is written.

The book is widely available through academic publishers and secondhand markets. You do not need the latest edition for exam purposes since the core material has not changed in decades. Earlier editions are functionally identical for what most people need. One last thing. When you get to the immunization section, pay attention to the difference between single-liability and multi-liability portfolios. Kellison presents the single case cleanly. The multi-case is where assumptions break down in the real world because you cannot perfectly match duration and convexity across multiple cash flow streams with a finite set of assets. I learned that the hard way on a project where the portfolio kept drifting out of immunization after the first rate move. Rebalancing frequency became the real problem, not the theory. If you want a supplement, combine Kellison with practice problems from the SOA review manuals. The theory alone will not carry you through the exam or the job. The problems do the carrying.