Diversification : Investor Level or Project Level?


Recently I read a book on capital budgeting and the book will say that the company needs  to diversify its project investment portfolio similar to what is suggested by Modern Portfolio Theory (MPT). It is said that

…. In a world without uncertainty, an investor would choose to own the investment that will provide the highest return over their holding period. In the real world, fraught with uncertainty, investors cannot possibly know which investment will provide the highest return, or even if an investment will earn a profit. Therefore, it makes sense to spread one’s funds across several investments in the hope that some will be profitable enough to more than offset the losses of others. This is known as diversification

I don’t really concur with the way that book tries to convince to use or apply MPT to project investment portfolio, yet my personal opinion:

I see that project portfolio and stock portfolio is certainly different stuff. Reducing risk in stock portfolio by applying diversification is well documented, pioneered by Dr. Markowitz in his ground-breaking paper, leading to the birth of MPT.

My question, can we bring the MPT concept to Project Investment Portfolio?

My personal opinion, no, we can’t. Though logically, it makes sense to not put all your eggs in one basket, yet, project is different in many fronts compared to stock. The specific risk is more dominant, and in many cases, the company might not have a chance to remove it at all by investing in many more projects both in the same industry and/or different industries. It might even make the company’s business risk higher, if it invests its money in the project that is beyond its expertise.

Additionally, project investment is difficult to be compared to stock investment. The specific project risk under different sectors or industries (if this is what the diversification under MPT means) might even make the company’s risk higher, for example, investing in project that is beyond the company’s expertise or industry. Additionally, switching is almost impossible to do once the company puts in the money. The issue with this project portfolio, the specific risk plays much bigger role compared to the stock portfolio, and in many cases, probably is not easy to remove at all.

Respondent 1 to Karnen:

The risk diversification assumes that you are constraint in one dimension: money.

However, in reality, companies are constraint in money and other aspects like time, management attention, specialization etc.

So I think that diversifying projects like portfolios is not a practical exercise since the company will not be able to manage it. However all else equal – it is better to invest in low correlation projects than in high correlation.

Respondent 2 to Karnen:

More generally, diversification matters at the investor level.  The firm doesn’t need to diversify, since its investors can. The potential gain from internal diversification is that it may support more leverage.

Respondent 3 to Karnen:

Well, there are mixed feelings and special considerations….

A firm could diversify if it sells/produces products/services that have “inverse” behaviors… For instance, imagine a firm with different products that sell in different seasons of the year. It’s a kind of diversification… Or a firm that produces the same product BUT has customers in different areas of the same countries with peaks at different dates. In fact, I remember I worked in a printing firm that used to sell notebooks for schools. In Colombia we have two calendars: schools that start classes at February and schools that start say in July. The firm had diversified according to markets. Moreover, this firm diversified its products/services because it doesn’t only manufactured notebooks but (in that time 1960’s/1970’s) there were no computers and the firms used to keep the bookkeeping in physical accounting books. Yes, the firm manufactured those accounting books. But also, it printed calendars with coloured photos, and printed poster according to its customers’ needs… and so on. The firm was diversified….

Agree, investor in equities are more prone to diversify, yes. And Markovitz theory was developed for those investors. No discussion on that. And no discussion on flexibility to change portfolios when it is based on stocks, say…. However, for a firm it is good to have some degree of diversification and yes, it is not easy to switch from one given portfolio to another….

Karnen to Respondent 3:

I guess, diversification is an old stuff, long before MPT was introduced (mathematically) by Harry Markowitz. This diversification makes sense since we don’t want to put all eggs in one basket. No discussion about that.

Yet, my big question: can we use MPT to apply practically to project diversification. I believe even Mr. Markowitz doesn’t want to bring his proposition to project level. Mr. Markowitz uses stock portfolios, and built that efficient stock portfolio frontier. Will that be possible again to bring that MPT and its efficient frontier to project diversification?

I just copied below from Principles of Corporate Finance textbook by Brealey, Myers and Allen (12th Edition, McGraw-Hill Education, 2017, page 184). This brings me home to this understanding.

We have seen that diversification reduces risk and, therefore, makes sense for investors. But does it also make sense for the firm? Is a diversified firm more attractive to investors than an undiversified one? If it is, we have an extremely disturbing result. If diversification is an appropriate corporate objective, each project has to be analyzed as a potential addition to the firm’s portfolio of assets. The value of the diversified package would be greater than the sum of the parts. So present values would no longer add.

Diversification is undoubtedly a good thing, but that does not mean that firms should practiceit. If investors were not able to hold a large number of securities, then they might want firms to diversify for them. But investors can diversify. In many ways they can do so more easily than firms. Individuals can invest in the steel industry this week and pull out next week. A firm cannot do that. To be sure, the individual would have to pay brokerage fees on the purchase and sale of steel company shares, but think of the time and expense for a firm to acquire a steel company or to start up a new steel-making operation. You can probably see where we are heading. If investors can diversify on their own account, they will not pay any extra for firms that diversify. And if they have a sufficiently wide choice of securities, they will not pay any less because they are unable to invest separately in each factory. Therefore, in countries like the United States, which have large and competitive capital markets, diversification does not add to a firm’s value or subtract from it. The total value is the sum of its parts.

This conclusion is important for corporate finance, because it justifies adding present values.

The concept of value additivity is so important that we will give a formal definition of it. If the capital market establishes a value PV(A) for asset A and PV(B) for B, the market value of a firm that holds only these two assets is

PV(AB) = PV(A) + PV(B)

A three-asset firm combining assets A, B, and C would be worth PV(ABC) = PV(A) + PV(B) + PV(C), and so on for any number of assets.

We have relied on intuitive arguments for value additivity. But the concept is a general one that can be proved formally by several different routes. The concept seems to be widely accepted, for thousands of managers add thousands of present values daily, usually without thinking about it.

As Brealey, Myers and Allen said above, if shareholders could do the diversification by themselves, then they will only pay for the market risk of the company’s business that they can’t diversify away. Any idiosyncratic nature from the unsystematic risk being taken by the company in any projects will be not be compensated by the investors.

The company of course could put their money into many projects that, is expected to have low correlation between one project to another project, or even put their projects in different geography or even different segment of business. However though logically by doing the aforementioned, we expect that the risk will be reduced from the diversification, yet behind this way of thinking, the company might have another risk coming in to counter that expected lower risk from diversification. The more projects in different segment, geography, etc. might give the company a challenge to manage such business and whether the management has such skill and business knowledge to deal with the conglomeration of the business. Not many business could have such skills and expertise. However, it might be whatever the risk adding to the project portfolio, it could be specific project risk, which in nature, non-systematic risk, and the investor could easily diversify it away using the portfolio diversification.

Respondent 4 to Karnen

But sometimes I really like the idea of Markowitz.  In the extreme case you can invest in every asset in the whole economy and then you should earn the overall growth of the economy plus the amount transferred from non asset owners to owners.

Brealey and Myers should simply go back to Merton Miller.  Miller said that leverage did not matter because investors could create leverage themselves.  Of course investors can diversify themselves so what value is added by companies doing something that investors can do anyway by themselves.

I think about the diversification issue in terms of venture capital and high tech companies.  They try to earn about 30% on start-ups that have demonstrated proven concepts and about 50% with no proof of concept (this is what private equity apparently wants).  Maybe on average they earn something like 10% or even less.  This means they are making a whole bunch of bets to get a few successes.  This is a kind of diversification that seems to promote innovation and maybe without this diversification there would be less investment in start-ups.

Karnen to Respondent 4:

Yes, I heard that 10% success rate from one of the articles or books about start-ups, this is why the expected [required} rate for start-up or any seed stage of entrepreneurship is quite high, even reaches 50%. However, this makes finance sense.

About the diversification, I guess, if the company’s owners could do the diversification by themselves, then logically (at least from MPT), no point for the company to do the diversification at its project level. If the owners do not have such opportunity for diversification, for instance, many privately owned corporation, then diversification at project level might make sense. Having said that, it doesn’t mean that diversification at project level is not needed, it is still necessary to  reduce the risk level, leading to higher leverage capacity for the company overall, in this case, the company could assume higher debt-to-equity ratio.

Another thing about this diversification at investor level vs project level, I guess, related to the size of the money being invested and agency issue.

Let’s say we flip a coin 100 times for a dollar bet each time. Investor might accept this. However at project level, the management of the company might not have such thing, they might flip a coin once for $100, so either they get that money back or lose it all (note: this assumed bell-shape probability distribution, though in reality, we should have this distribution is skewed to the right).

Agency issue also comes up since disproportionate share of their potential future (compensation) is tied up to that project’s success as well.

What I am trying to say, MPT cannot be brought to project level diversification discussion. Stock portfolio and project is again two different stuffs.

Respondent 5 to Karnen:

I disagree. An investment is an investment. The idea of diversification and optimization is what is important. Stocks were used in the chapter simply because they are familiar and we have relatively continuous data on returns which makes it easy to measure risk. Would you argue that a property insurance company should concentrate all of its policies in one small geographic area? I wouldn’t as they could be bankrupted by one natural disaster in that area. Geographic diversification seems like a good idea. Just a couple of weeks ago I saw an academic paper using Markowitz’s ideas in marketing. Years ago, I recall seeing an application in information technology projects. A quick Google search turned up this article:


You might also do a search on “project portfolio optimization,” and keep in mind that it is the general principal, not the exact details that matter.


Neil Steiz and Mitch Ellison (Capital Budgeting and Long-Term Financing Decisions. 3rd Edition. The Dryden Press. 1999. Page 417) said that the problems encountered by managers in applying mean-variance MPT include the following:

  • Indivisibility of Assets

Full mean-variance portfolio analysis is based on the assumption that securities are infinitely divisible, while capital investments often come in very large, indivisible units…. Fortunately, the lack of divisibility can be handled with integer quadratic programming, which forces the selection of only whole units of specified investments. Constraints can also be used to limit the choice between zero and one unit.

  • Holding Period Choice

We could use mean-variance analysis to find the portfolio of capital investments that has the highest net present value for each level of risk, where risk is measured as the standard deviation of the net present value. We could do the same thing with the profitability index or the internal rate of return. Unfortunately, though, the efficient frontiers that result are difficult to interpret because asset lives are not all identical. (Note from Karnen: MPT is about stock that has not limited life. The stock of that company could be delisted, yet, the company itself might still exist, via merger, acquisition, etc., unless it goes bankrupt or be liquidated by the owners)….To avoid ambiguity caused by unequal lives, we usually measure return over some holding period. One alternative is to use a long holding period, such as the life of the longest-life asset being considered. To measure portfolio risk and return over this holding period, though, it is necessary to identify all assets that will be selected between now and the end of the holding period. In addition, there is no date at which the lives of all assets to be acquired both now and later will end, unless the company plans to liquidate. Therefore, some terminal value estimates will be required as well. Beyond these difficulties of application, there is a serious conceptual problem. Risk analysis based on the results over a multiyear holding period does not give a complete picture of risk because it ignores fluctuations during the holding period. Problems like the risk of not being able to meet financial obligations during a bad year are missed when a long holding period is used, so the long holding period gives an incomplete picture of risk. …..These problems are avoided by using a short holding period, such as a year, and using the present value of remaining cash flows as the value of each asset at the end of the holding period. Present value of remaining cash flows would be used instead of market value because many profitable capital investments cannot really be sold after such a short holding period, without suffering a large loss. Unfortunately, the lack of opportunity to sell an asset for the present value of tis cash flows limits our ability to modify the portfolio at the end of the holding period. This disadvantage is not assigned a specific value by the model, so the short holding period also gives an incomplete picture of risk and return.

  • Defining the Borrowing-Lending Rate

[Karnen: Once a risk-free asset is introduced and assuming that investors can borrow and lend at the risk-free rate , the conclusion of Markowitz MPT,  and every combination of
the risk-free asset and the Markowitz efficient portfolio M is shown on the capital market line (CML) below.

Source: Equity Portfolio Management. Frank J. Fabozzi and James L. Grant. 2001. Wiley. Page 27.

The interest rate depends on the maturity of the debt, and the only risk-free interest rate for a holding period is that for debt with a maturity equal to the holding period. For a long holding period, the rate on long-term bonds can be used. For a short holding period, a short-term rate would be appropriate. However, the possibility of entering into long-term loan agreements must be recognized, even if a short holding is used for the analysis. One way to recognize the long-term borrowing is to treat long-term debt as a risky asset in which a negative position can be taken. The long-term debt is risky over a short holding period because the terminal value of that asset will depend on the interest rate prevailing at the end of the holding period.

  • Data Needs

Another problem with the use of mean-variance portfolio analysis for capital budgeting is the tremendous amount of data required. Covariance estimates are needed for each asset in relation to every other asset. The number of covariance terms needed (excluding variances which are covariances of an asset with itself) is 0.5N^2 – 0.5N where N is the number of assets being considered. Including expected return and standard deviation for each asset, the total number of terms needed is 0.5N^2+1.5N. If 1,000 capital investments are being considered, 501,500 risk and return measures must be estimated. The estimation is made more difficult by the fact that covariances and correlation coefficients are not the types of measures for which the average manager will be prepared to make judgmental estimates.


Most of the problems mentioned above are related to the assumptions of MPT.

MPT is built from the perspective of well-diversified shareholders. If shareholders are well diversified in their own stock portfolios, then following MPT, their main concern is only about that asset’s contribution to non-diversifiable or systematic risk for a broadly based portfolio, and in this case mean-variance analysis of the company’s portfolio is not needed for that purpose.

Applying MPT to capital budgeting or project level, this will mean that we analyze the risk from the perspective of the company and its manager.

(to be continued with more arguments to say that diversification at the project level could reduce risk but create no VALUE to investors that could diversify by themselves.)


Leave a Reply

This site uses Akismet to reduce spam. Learn how your comment data is processed.