Aquisitions vs. Buyback
Board: Berkshire Hathaway
The following is my attempt to answer, with some numbers, the question:" When may stock buybacks be a better idea for BRK than acquisitions."
The Annual Report presents the growth rate of the components of the 2 column metric. We will get an update in a few weeks. The rates in the 2012 AR were 19.4% for Investments and 20.8% for Earnings of the non-insurance subsidiaries. These rates are the annual compounded growth rates since 1970.The 20% trailing rate for the subsidiaries' earnings is indeed impressive. This large growth rate is attributable to BRK's acquisition strategy, as the organic growth rate of the subsidiaries' earnings is much less than the 20% rate. But, this long-term growth rate is also dropping over time. As the size of the subsidiary base grows, it becomes more difficult to sustain a high growth rate in subsidiary earnings both because bigger and bigger acquisitions are needed, and the average amount of money available for acquisitions, as a fraction of the size of the value of the acquisition base, diminishes. Roughly speaking, excluding borrowings, the amount of new money available for acquisitions grows by BRK's total annual net operating earnings, currently about $15B/year. $15B is about 10% of the current value of the subsidiaries, so, purchasing equally good companies can provide about a 10% average boost to earnings and the rest of the annual subsidiary earnings growth has to come from organic growth. Organic growth would have to exceed 10% for total growth to reach 20% and this is and will continue to be harder and harder to achieve, since the 10% acquisition growth rate number will probably slowly trend down, while the 10% organic growth rate is already on the high side. For example, the AR lists marginal growth rate of column 2 earnings for 2012 as 15.7%.
Given the above context, I decided to build a little model that could be used to evaluate the trade-off between continued acquisitions versus buying back BRK shares at varying discounts to Intrinsic Value (IV) and at varying ability to find high growth rate and sizable acquisitions. The model was pretty simplistic, and therefore fraught with some assumptions open to debate, but I tried to do it in a way that the assumption variances/errors would affect both options, acquisitions vs.buybacks, in the same way and thus not bear directly on their relative comparison, although they would have more of an effect on any absolute prediction of BRK's future IV. The metric I used to compare the 2 options was BRK's predicted (differences in) per-share IV five years in the future. I took my current estimate of the 2 column metric as a proxy for IV for both options and assumed $50B available for either new acquisitions or buybacks. I won't try to present the rest of the details here, but rather jump to the results. The 2 major findings,using what I judged to be nominal (i.e., neither too optimistic nor too pessimistic) values for the key parameters, are:
1. An acquisition strategy results in higher per share IV 5 years out than a buyback strategy, if the subsidiary earnings can continue to grow at 20% and if buybacks are done at a buyback price below 1.2Book/share .
2. A buyback strategy works better if subsidiary earnings can be only grown at an annual rate of 15%, as long as buybacks are done at a price below 1.4Book/share.
A few observations I drew from this analysis are:
1. The current 1.2Book buyback threshold may not be as arbitrary as it appears to some, as it is more or less an equivalent economic alternative to the recent historical acquisition strategy in which the 20% subsidiary earnings growth rate was achieved. So, maybe WEB had this in mind when going with 1.2Book.
2. As it gets more and more difficult to achieve the 20% subsidiary earnings growth rate, I expect the IV growth equivalent buyback threshold, as a % of Book value/share, to increase. Also, as Rational has explained, the IV/Book ratio is probably growing over time and this means the 1.4Book or whatever other economically neutral number BRK is using at the time will slowly drift upward.
3. However, it may still be hard to do large buybacks at a favorable price in the open market. Remember, my model did $50B over a five year period, an average of $10B/year. But, if we really are approaching a point in which buybacks at say 1.4Book are competitive, what can be achieved in the way of buybacks in a practical/realistic way going forward? Perhaps tender offers at an attractive premium to the then market price would be a way forward and more manageable than waiting for old rich shareholders to die.
The 1.2Book to 1.4Book numbers are somewhat sensitive to assumptions, but not dramatically so. The trade-off effect appears to me to be pretty robust--as subsidiary earnings growth potential diminishes, buybacks at increasing prices become more attractive.
I then changed the model from looking at spending $50B over 5 years on either acquisitions or buybacks to spending $15B for 1 year, i.e., basically putting all the operating earnings into acquisitions or buybacks each year. At this level, the upper limit on the attractive buyback price rises to about 1.5Book/share when subsidiary earnings growth is projected at 15%/year, whereas at 20%/year breakeven is about 1.2Book.
The economic trade-off between the 2 options is also simple enough that the BRK Board could easily utilize it after WEB is gone, so I don't assume dividends are inevitable, at least for a long time.
I don't know if or how quickly WEB will place an increased weight on buybacks. Nor is the issue a complete either/or. Acquisitions of strong businesses at a good price could still be done, when they are available, while still doing more buybacks. Maybe we will see an increased emphasis on buybacks shortly, or maybe WEB is waiting for the next bear market to buy big quality at good prices. When such are not available, buybacks can act as the place to deploy excess cash. To me, this would be much better than instituting a dividend policy, under either WEB or his successor. My own sense is that increased buybacks will start sooner rather than later, since the economics are pretty compelling and WEB can no doubt do a more precise, higher-quality assessment and has better market judgement than me. And, under these conditions buybacks can be justified by their economic logic alone, not by a desire to support the stock price in the short term, a nice rational argument for a very rational WEB/BRK. We know IBM's buyback strategy played a key role in WEB's decision to invest in it.
Perhaps turning the buyback knob will be BRK's response in the Annual Report or Annual Meeting to a failure of the 5-yeat test. I would be surprised if it takes more than 5 years to put a much bigger emphasis on buybacks, but I may be constrained in my thinking by the limitations of my analysis and my view of what the market will serve up as buying opportunities. Of course, if BRK's market price climbs to IV, BRK will have to find a different solution to its excess cash problem, but that will be a nice problem to have.