Every year, I reread fundamental valuation literature, particularly Revenue Ruling 59-60 and the newest Ibbotson SBBI Valuation Yearbook.  After almost 30 years of doing this, I still pick up little things that make my reports better!

1. 59-60: I now break the 8 factors to consider into 10, separating the economic and industry outlooks and the size of the interest and prior (company stock) sales.  No big deal, but it just reads better!
2. SBBI: this year I realized that my general practice of raising the equity cash flow discount rate 5% in the terminal year (above that for years 1-5 to account for higher long-term risk) is just a variant of the two-stage growth model.  Again, no big deal, but I feel better about doing so!

It strikes me that the Crash of 2008 could provide a way to further inform our knowledge of and ability to quantify lack of marketability discounts based on evidence from markets other than the public markets.  After all, restricted stock and pre-IPO studies are based at least in part on liquid, publicly traded prices.

By contrast, the mortgage backed securities market, which became almost entirely illiquid when the pool of willing buyers almost evaporated, resembles the normal conditions of the market for private equity minority interests, for which there is no market data.

I am not sufficiently knowledgeable of these markets or their data sources to do more than pose this as a question.  Maybe this will inspire someone else who has the knowledge!

If you have read these posts for a few years, you know that I strongly prefer to value equity interests based on cash flow to them.  This lets me directly calculate and demonstrate cash-on-cash return on investment (the third part of the justification of purchase test).  It also lets me sidestep using WACC, because I explicitly forecast the level of debt each year.

My financial forecasting model is probably just like yours.  Mechanically, we iteratively input the projected level of (short- and long-term) debt each year until cash (the residual that makes the balance sheet balance) is zero or some specified minimum / maximum value.  It is the last assumption we input.

If we do it this way, we are making an important implicit assumption: the policy is to minimize debt each year.  That is one of many possible strategies:

1. The business could be debt-free in the last historical year, might not require any debt in the future, and the control owner may be debt-averse (even though prudent borrowing could increase return on equity, if the cost of debt is less than return on assets).  In other words, the business will remain debt-free.
2. The business may carry some debt or need some in the future, but the control owner wants to minimize it.  This is the strategy outlined in the second paragraph of this post.
3. Same as (2), but the control owner wants to maximize it in the short run.  By this, I mean the business will borrow as much as it can using an asset-based facility (e.g., 80% of receivables under 60 days, etc.).
4. Same as (2), but the control owner wants to maximize it in the long run.  By this, I mean the business will borrow as much as it can using both an asset-based facility and term loans supported by fixed assets and cash flow.

These four strategies cover the extremes of from no to maximum debt.  Three other possible ones fall somewhere in between the extremes, depending on case facts:

1. The business has debt but pays it down as scheduled.
2. The business borrows at historical industry averages (say of debt to equity).  The problem here is determining whether historical averages will apply to the future and whether they are sensible for the subject.
3. Debt levels as specified by the control owner (assuming they are feasible).

In all strategies, we have to consider the control owner’s plans and preferences; they, not the minority owners, decide debt policy.  If additional equity capital is going to be needed, we have to consider that as well.

This looks a lot more complicated than it really is!  A few on-point questions answered by the control owner will usually narrow down the possible strategies to one or maybe two.  In my experience, most business owners opt for strategies 1, 2, or 5, and few (except for real estate owners) opt for 3 or 4.  I have never encountered an owner who follows strategy 6, and only a few who follow 7.

A colleague had a great question: how do you decide which approaches and methods apply to an appraisal?

My first answer – it depends on case facts and circumstances and data availability – is true but not particularly helpful.  After I thought about it for a while, I came up with what I hope is a better one.

To keep it simple, restrict the discussion to operating companies; exclude holding companies (the latter are usually valued using the Asset Approach, although assets such as stock in subsidiaries might be valued using any approach).

A life cycle analogy is apt: birth, toddler, adolescent, adult, middle age, old age, and death.  Except for birth and death, it is hard to define the boundaries between them with precision, but most people would probably not differ by more than one stage in their assessment of where a person is in their life cycle, using the word “potential” for their future earning power.

1. A newborn infant’s potential is unknowable.
2. A toddler’s potential is highly uncertain.
3. An adolescent’s potential is starting to become visible.
4. A young adult’s potential is visible.
5. A middle-aged person’s potential is extremely visible.
6. An old person’s potential is starting to decline.
7. A dead person’s potential is zero (unless heaven has employment opportunities).

Extend that analogy to operating companies:

1. A newborn company’s earning power is zero.  We value it based on the investment made in it (Asset Approach).
2. An early-stage (toddler) development company’s earning power is not quantifiable; we use option models to value it.
3. A later-stage (adolescent) development company’s earning power can be guessed at; we use the probability-weighted expected return method to value it.
4. An established (young adult) company’s earning power can be projected using a DCF methodology (varying annual growth rates).
5. A long-established (middle-aged) company’s earning power can be projected using a SPCM methodology (stable growth rate).
6. An old-line (old person) company’s declining earning power can be projected using a DCF or SPCM methodology (varying or stable decline rates).
7. A dying (dead) company’s earning power is less than its liquidated value.

It is difficult, if not impossible, to apply the Market Approach (absent same-company stock transactions) to the first two and  the last one.

There is judgment as to what stage of the life cycle a company is in, but in my estimation, most appraisers would not differ by more than one of them.

Most of the time, I provide limited calculations of value when entire companies or fractional interests are being bought / sold, the parties are unrelated, and there is no prospect of litigation.  Although I could, with proper disclosure, be non-objective (favoring my client), I much prefer to be objective, tempering both my client’s (perhaps unrealistic) expectations and at the same time creating credibility for the counterparty to the transaction.

In a non-synergistic transaction (where the buyer is financial), I provide a range of reasonable values based on reasonable assumptions.  In a synergistic transaction (where the buyer is strategic), I provide a range of reasonable values based on stand-alone value (zero synergy) at the low end and strategic value (full synergy) at the high end.  This gives the parties a quantified range of values – a ballpark – within which to negotiate based on whatever assumptions they can agree upon, which is more constructive than uninformed “hard bargaining.”

For a fixed fee, I will run as many cases as are needed for my client (and, if asked, the counterparty) to become comfortable that the range is reasonable.  The incremental time to revise a basic model to reflect new assumptions is small, and this facilitates “what-if” questions without the client worrying that the fee meter is running.  As part of this, I suggest that the client show my work to the counterparty to prove that the value range has been thought through.

The vast majority of these calculations use the Income Approach with a justification of purchase test.  This clearly demonstrates the effect and sensitivity of various assumptions.   It also demonstrates the cash-on-cash return on investment the buyer will realize.  The Market Approach does not allow for this: all one can do is vary the price multiple, but how does one tie that quantitatively to different assumptions?

I have found this to be very transparent to the parties, with all assumptions clearly disclosed and explained, the calculation methodology explicit, and the results justified.

I do not advocate for my client or help them negotiate: their transaction advisors are better equipped to do that, and my independence is not compromised.  It is up to the parties to agree on the assumptions; in essence, I am just a calculator facilitating their exploration of their impact on value and an independent, objective voice of reason.

Excess cash is the amount distributable to shareholders without affecting company risk and return.  If it is distributed, cash and equity value decline dollar for dollar; nothing else changes.  Equity value declines by the distribution, and owner cash increases by the same amount.

How can we quantify excess cash?  One way is to use industry norms (like RMA data) to determine cash as a percentage of assets and apply that to the subject to compute the excess.  In essence, this is a rule of thumb: the problem is that no two thumbs are necessarily alike.

In my experience, case facts and circumstances like these make rules of thumb dangerous:

1. If the company’s balance sheet ratios differ greatly from the industry average, chances are its (excess) cash will also differ greatly.
2. Loan covenants may require a certain minimum cash balance.
3. Companies often have cash management systems that allow them to operate with no cash and a line of credit.
4. Some companies operate with permanently negative working capital (current liabilities exceed current assets).  Barbershops, for example, do not have receivables and carry very little inventory, but may have (large) payables.
5. All of the cash held by a debt-free company that is (highly) cash flow positive is probably excess.
6. A company saving for a future outlay (such as a big distribution or a capital investment) may have a large cash balance today that will not be excess after the outlay is made.
7. The cash of a company in the process of selling assets may also be excess.

Try to avoid rules of thumb for excess cash!

Simple questions can be very thought provoking!  An attorney e-mailed me, asking whether Revenue Ruling 59-60 is applicable to investment value appraisals (in this case, one involving a strategic acquirer).  I am glad that she did not call, because my answer, upon reflection, was different from my initial one!

Obviously, the Revenue Ruling 59-60 definition of fair market value does not apply to investment value.  The “Factors to Consider” (Section 4), however, are very useful:

1. Nature and history of business:  a strategic acquisition will create some sort of value-adding synergy (raising revenue, reducing cost, and / or reducing risk), so historical results will not be very meaningful, and it is essential to assess how the nature of the acquired business will change.
2. Economic and industry outlooks: if the synergies are material, these effects will probably overshadow economic and industry trends in the short run, while the acquisition is integrated, unless the economic and industry trends are material.  The latter will be more significant in determining long-term sustainable growth rates.
3. Book value and financial condition: will probably not be relevant for valuation purposes because (depreciated) historical cost will bear no relationship to investment value, and unless the business is bought for liquidated value, the latter premise is irrelevant.  Financial condition may change materially due to the acquisition financing structure.
4. Earnings and dividend capacities: will probably be paramount, along with point 1 above.
5. Goodwill and intangibles: will come into play if a purchase price allocation is needed.
6. Past sales: will probably not be relevant because of point 1.
7. Interest size: depending on whether discounts and premiums apply (this is usually negotiated), this may be relevant.
8. Comparables: in my experience, comparables are not very helpful in investment value determinations because it is unknown what synergies were inherent in comparable (strategic) transactions.

The bottom line is that roughly half of the Revenue Ruling 59-60 considerations are relevant!

One thing that it does not address regarding investment value is the relative bargaining position of the parties, which will affect how synergies are shared between buyer and seller.  A buyer in weak position (e.g. their business is unprofitable on a stand – alone basis) will have less leverage than one whose business and financial positions are very strong.

This post generalizes the preceding one (on Monte Carlo Simulation).  Regardless of what we are appraising (i.e. the final value or any of its components), we can use two methodologies: discrete or probabilistic.

Discrete methods use point estimates for assumptions (inputs) and yield them for results (output):  if revenue is $100 and cost of goods sold is $40, then gross profit will be $60.  Nothing is stated or inferred about the probability distribution of revenue, cost of goods sold, or gross profit.

Probabilistic methods use probability distributions for assumptions and yield them for results.  Parameters (such as the mean or standard deviation) characterize some probability distributions, while others are specified (e.g. 90% chance that sales will be less than $100, 60% chance that they will be less than $50, etc.).  Here, probability distributions are explicit for the inputs and the output.

In simulation, each input (and output) is a probability distribution.  In regression, the inputs are point estimates (such as price and sales in the Direct Market Data Method) for which probability distributions and their parameters are calculated along with that of the output.

Some methods have both discrete and probabilistic aspects: probability-weighted scenarios (e.g. “best case – worst case”) for forecasting financial results and Black-Scholes option models (where volatility, a statistical measure, is part of the valuation formula).

To summarize, discrete methods use and yield point estimates but include nothing about probabilities.  Probabilistic methods assume and yield probability distributions but imply nothing about point estimates.

So how does one decide which methodology to use?  I believe that it depends on:

1. The quality and supportability of the input data: how well can you defend an assumed probability distribution?
2. The nature of the assignment: is a point estimate required (e.g. estate taxation) or is a probability distribution more useful (in an actual transaction)?

Remember: there is no way to convert a pure discrete result into a probability distribution, and vice versa.  Discrete methods say nothing about probabilities, and probability distributions say nothing about the likelihood of a point estimate.

In most appraisals, we use deterministic models whose assumptions (inputs) and results (outputs) are single numbers.  For example, if we are modeling gross profit as sales minus cost of goods sold, we assume sales to be $100 and cost of goods sold to be $40, with the result that gross profit is $60.

Monte Carlo Simulation (MCS) is a probabilistic model whose assumptions and results are probability distributions.  In the above example, we might assume sales to have a mean of $100 and a standard deviation of $10 (a normal probability distribution) while cost of goods sold could be from $35 to $45 with any value in that range having equal probability (a range distribution).  (Good software allows you to assume any desired kind of probability distribution.)  The result will be a probability distribution of gross profit.  I did not run the numbers, but in the example the distribution might be characterized by a mean and standard deviation (normal) or it might be something else.  Regardless, we will be able to make statements such as “there is a 10% chance that gross profit will be less than X$.”  Nice tables and graphs can summarize the output distribution.

MCS is a generally accepted technique, but it poses three problems to which there are no general solutions:

1. How do you specify and support the input distributions?  We often have small sample sizes (e.g. five years of relevant historical sales and cost of sales), which reduces statistical confidence.  Our or management’s subjective judgments will be highly significant and often hard to defend.
2. How do you account for correlations (relationships) between the input distributions?  If cost of sales declines with sales we have to be sure to model that, which is not hard.  If, however, we have manually made “plug” assumptions (such as adjusting debt to be sure that cash is $0 or some stipulated amount), it is not clear that MCS models can handle this.  (For more on this, see my January 16, 2016 post entitled “Copulas.”)
3. How do you condense the output distribution into a single number – which we have to do for taxation, financial reporting, and many other assignments?  What if the “most probable” output result has a range of values?

My take on all this is that MCS is a helpful tool for explicitly modeling uncertainty, but it rests on potentially troublesome input assumptions, logical problems (copulas), and the difficulty of reducing an output distribution to a single number.

MCS can be VERY helpful in actual transactions, where buyers and sellers can deal with uncertainty (by techniques such as earn outs), but its usefulness for point estimate assignments (like taxation) is limited by its challenges.

Do you ever wake up in the middle of the night with weird BV thoughts?  If not, I am happy for you.  If so, welcome to the club!

My most recent weird night BV thought was whether the size premium component of the equity cash flow discount rate applies to passive holding entities.  (Believe me, I have had even weirder ones, none of which I am going to share, in order not to scare you.)

Up until this thought, I always automatically applied a size premium whenever I calculated the cost of equity capital, regardless of whether the subject was an operating company or a passive holding company, and to be honest, the question had never entered my mind (such as it is).

The specific question that awoke me was whether there is a difference in risk between a $100,000 portfolio and a $100,000,000 portfolio.  At first blush, my answer was that there is none, because both can invest in (say) a stock index fund and have the same risk / return profile.  On that logic, there would be no rationale for a size premium.

I could not find any discussion of this point in the literature, so I put the question to my American Business Appraisers National Network (businessval.com) colleagues.  They always provide great advice, and this was no exception.  Brandi Ruffalo pointed out that the $100,000,000 investor has access to a wider range of investments, hedging tools, institutional research, and sophisticated management than the $100,000 investor has.  The smaller portfolio will therefore be more risky, justifying a size premium.  Sherri Smith mentioned Berkshire Hathaway as an example of a holding company that is less risky than a smaller portfolio.  Jim Lurie pointed out that each portfolio investment has a size premium, and these could be calculated and weighted-averaged for the portfolio.

The moral of the story is when in doubt, ask your colleagues!  Do NOT call THEM in the middle of the night, but feel free to call ME THEN because I will probably be awake pondering some new weird BV thought!