96-04-12 rizzo@hogpa.ho.att.com (Anthony R Rizzo)
The following is a work in progress. I'm still struggling with it. Therefore, I would greatly appreciate your comments, observations, and suggestions for improvement, even more than I usually do.As you read it, please keep in mind that the targeted audience consists of people in the non TOC world.
Flush - The Investors' Value Meter
by Tony RizzoHave you ever wondered why the stock markets appear to be so illogical? Doesn't it seem odd that the price of a company's stock should rise, when all logic says that it should not? Yet it happens, frequently.
This seemingly illogical effect can mean only one thing. Investors are responding to a measure, of the companies in which they invest, that does not give them an accurate reading of the real worth of those companies. In other words, the investors' value meter is badly malfunctioning. But if the value meter that most investors use is malfunctioning, then what is a reliable value meter? The answer to this question is provided by the Theory of Constraints, and it is the subject of this article.
The Theory of Constraints is a philosophy that treats a corporation as a complete and complex system, rather than treating it as a collection of non interacting components. To illustrate the significance of this statement, let's use a simple analogy. If the many departments of a corporation were chain links, then the Theory of Constraints would tell us to view the corporation as a chain.
In reality, this chain analogy isn't far from the mark. Many of the processes employed by a corporation happen sequentially, though there may be some overlap. For example, marketing must determine what product the company should sell, before engineering can design it. Engineering must design the product, before manufacturing can make it. Manufacturing must make it, before the distribution department can bring the product to stores. Distribution must ship the product, before the stores can sell it. So, if many of the processes that are employed by corporations do happen sequentially, doesn't it make good sense to focus on how those processes interact? It makes perfect sense, and that's exactly what the Theory of Constraints tells us to do.
A for-profit corporation is a system designed to generate profits. Like many other systems, it is a complex system. Like the chain in our analogy, it has only one constraint at any one time. What is a constraint? A constraint is anything that prevents the system from achieving more of what it was designed to achieve. A chain is designed to transmit force. For a chain, its constraint is its weakest link. That's what prevents the chain from transmitting larger and larger forces. For the profit-generating system that is a corporation, its constraint is whatever prevents it from generating more profits for its owners.
This is important, because, if management is to really strengthen the system's ability to do its job (generate profits), then it must focus on the weakest link. It must focus on the constraint. Just as strengthening an already strong link in a chain does nothing to strengthen the chain, so too, improving anything in a corporation that is not the constraint does nothing to increase profits substantially. Such so-called improvements have no substantive bottom-line impact, because the corporation's ability to generate profits is constrained (limited) by something else, which hasn't been improved. Here is an example that illustrates this important point.
More than occasionally, a company's ability to generate profits is restricted by the limited availability of a critical resource. Consider a manufacturing company with a hot product on its hands. Such a company could find its sales (and profits) limited by its ability to manufacture the product. The Cabbage-Patch-Doll craze of years past offers one such example. The company's manufacturing capacity, in turn, might be limited by a single resource in that product's manufacturing line. In that case, if we were to significantly improve everything else in the company, then we would see only a negative bottom-line impact. That single, remaining resource would continue to restrict the company's ability to manufacture and sell the product, while our extensive improvement efforts everywhere else would drain away earnings. This is an example where the constraint is physical, i.e., it's a resource. We refer to such a resource as a Critical Constraint Resource.
But, if improving all the operations of a corporation doesn't make sense, then what should we do? The Theory of Constraints offers a very useful approach for improving any system, including one designed specifically to generate profits. It is called the five-step process of ongoing improvement, and as we might expect, it is perfectly logical.
The first step of this five-step process of ongoing improvement is to identify the constraint. Until we know where the weak link is, our attempts at improvements are likely to achieve little more than an increase in costs. Improvements to operations that do not constrain the company's earnings do nothing to alleviate the constraint. Therefore, step-one must be to identify the critical constraint resource. Think of this as a focusing step.
Step number two also is perfectly logical. Once we identify the critical constraint resource, we exploit it to the hilt. This means that we waste nothing of it. After all, this is the one resource that limits the company's ability to generate profits for us. It makes perfect sense for us to exploit it completely. Wasting any aspect of a critical constraint resource has a direct, negative impact on earnings.
The third step is called the subordination step. What do we subordinate? We subordinate all other operations, everything. To what do we subordinate them? We subordinate them to our decision to fully exploit the critical constraint resource. That's right. If we are to fully exploit the critical constraint resource, then we cannot allow any other operation to take precedence. Otherwise, that other operation might prevent us from fully utilizing our critical resource. That could result only in a decrease in earnings.
The fourth step is to elevate the critical constraint resource. This means that we increase its capacity, somehow. If the constraint were a piece of machinery, then we might buy another, or we might rent another, or we might buy some of the excess capacity of our competitors. However we choose to do it, the end-result is that we have more of that critical constraint resource, which in turn increases our company's ability to generate money for us.
This brings us to the fifth step. This is the shortest step. Yet, it is the most difficult. Step five is to go back to step one and begin the process anew, to prevent our own inertia from halting our improvement process. This is the five-step process of ongoing improvement that is suggested by the Theory of Constraints. It is perfectly logical, once we accept the concept of a constraint. It is also perfectly effective, as corporations like Texas Instruments, Avery Denison, Bethlehem Steel, and Delta Airlines (to name a few) have learned.
Yes, this is all very interesting. But what does this have to do with our value meter? Consider this. If we look at the stock markets and at how we use them to generate profits, aren't we looking at a system? Indeed, if we look at any investment opportunity, aren't we looking at a system designed to generate earnings? On a somewhat high level, of course we are. Well, then, doesn't this system include resources? Certainly, the brokerage houses with which we deal are resources. Occasionally, they are also critical constraint resources, particularly in times of high demand for their services. Certainly, the exchanges themselves are resources. But are these the only resources?
No! Embedded within this high-level system there is another most important resource; it is our own critical constraint resource. What is this resource? Let's read on. For the sake of discussion, let's say that we have set a long-term financial goal for ourselves. We may want to provide for our children's future, or for our own retirement. Whatever the reason, we want to achieve a substantial financial goal in our lifetime, and we want to achieve it through investments. What are the resources at our disposal? The answer is simple. The only resources at our disposal are money and time. These are the things that we control directly, and we need both to achieve our financial goal. But is either of these resources more important than the other? This is a critical question.
Let's see. If we had all the money that we wanted, and if we had it right now, would time be important to us? Clearly, it would not. In truth, we would have achieved our financial goal already. So, time would be of little concern to us, at least so far as it might impact our ability to meet our financial goal. Similarly, if we knew that we would live forever, would it matter to us if today we had only one dollar? Very likely, it would not, so long as we had the things necessary for survival. If we knew that we would live forever, then we could afford to wait as long as necessary, to achieve our financial goal. However, and here is a dose of harsh reality, no one has infinite money or infinite time. Every one of us has but a very limited supply of these resources. Therefore, both are important to us, and we have no valid reason to treat either one as the more important of the two. Further, since most of us are mere investors, these are the only resources under our direct control. Therefore, by definition they are critical constraint resources.
We are faced with a conundrum. We have not one but two critical constraint resources, or so it would seem. Actually, together, these are really one resource. Together, they represent the growth potential of our investment capital during our limited time on this earth. It is the combination of our limited money and our limited time that is our real, critical constraint resource.
Now, we can draw from the work of the mathematicians just a bit. When faced with having to optimize two quantities simultaneously, mathematicians use a little trick. They multiply the two quantities, and they optimize the product of the two. We can do the same quite easily. We simply multiply the money that we invest by the time that the money is invested. The resulting product is called FLUSH (someday I'll ask Eli Goldratt where he gets these names, really).
Is this all that there is to it? Not quite. We're forgetting a most important part. As our investment yields real profits for us, we need to include these in our measurement as well. After all, we make investments only because we expect to get a return on our money, don't we?
I think that we need an example. Let's say that we invest one hundred dollars in a new project. Let's say, too, that we invest our money all at once. Then, one day after we made the investment, our FLUSH meter (if we had one) would read -100 dollar-days. After one week, our FLUSH meter would read -700 dollar-days. Why the negative values? Simple! We've invested in the project, but we haven't received any earnings yet. Therefore, our meter should read negative. When will it read positive? If we've made a wise investment, then it will read positive after an acceptably short interval. Unfortunately, many of our investments never generate profits at a rate necessary to achieve positive values of flush.
This is why FLUSH is such a useful measure. With it we can compare and render objective judgment on investment possibilities. Every time that we invest our money in something, for any finite period, we are committing our critical constraint resource. If we are to exploit our critical constraint resource -- recall that the second step of the five-step process of ongoing improvement is the exploitation step -- then we need a measure of how efficiently we are likely to utilize that resource with candidate investments. Exploiting our critical constraint resource (the growth potential of our investment capital) requires that we maximize the rate at which our projects release that resource, so that it can "process" more investments for us. Therefore, FLUSH is a perfectly logical measurement for us to use, if we accept the validity of the Theory of Constraints.
To properly use FLUSH for this purpose, we compare two things. First, we compare the time at which projects achieve zero flush. Second, we compare the amount of positive FLUSH that the candidate projects are likely to generate.
The time that a project needs to achieve zero FLUSH is important, because it is the time that it takes for us to recover completely our critical constraint resource. All other factors being equal, the project that achieves zero FLUSH before the others is the most desirable project.
In light of this statement, consider the implication of investing in projects that never achieve zero flush. Such projects literally rob our critical constraint resource of significant portions of its capacity. For this reason, any project that achieves zero FLUSH within a reasonable period of time is infinitely more desirable than one that never achieves zero FLUSH.
The second quantity that we compare is the amount of positive FLUSH that projects generate over the longer term. This positive FLUSH is how we elevate our critical constraint resource. Think of it, if FLUSH is a measure of our critical constraint resource, then achieving positive FLUSH means that we are actually increasing the capacity of our resource. If a project could achieve this for us within an acceptably short interval, shouldn't we want to invest in it? After all, if we increase the capacity of our critical constraint resource, don't we gain the ability to make even more progress toward of investment goal? In fact, we do.
Therefore, FLUSH, which is the product of money and time, and which measures our capital's capacity for growth (in light of our limited time on earth), lets us apply the second and the fourth steps of the TOC process of ongoing improvement to our investment system. What about the third step, subordination? Well, I haven't figured that one out yet. When I get there, I'll have material for another article.
96-04-15 tsellan@MNSi.Net (Timothy Sellan) (Subj: Flush)
No discussion on the principle of Flush would be complete without direct comparisons to other methods of calculating the time value of money. What is the difference between Flush and Net Present Value (NPV) or the Internal Rate of Return (IRR)? Where the time and money values of Flush reach zero sounds a lot like the break even point of a Break-Even Analysis (BEA). If we were to modify a BEA to reflect "Throughput per unit of time on the constraint", would this be about the same as Flush?
96-04-21 rizzo@hogpa.ho.att.com (Anthony R Rizzo)
No discussion on the principle of Flush would be complete without direct comparisons to other methods of calculating the time value of money. What is the difference between Flush and Net Present Value (NPV) or the Internal Rate of Return (IRR)? Where the time and money values of Flush reach zero sounds a lot like the break even point of a Break-Even Analysis (BEA). If we were to modify a BEA to reflect "Throughput per unit of time on the constraint", would this be about the same as Flush?Here's how my neurons fire currently. The flush calculation gives equal (mathematical) weight to both the time and the money component. The NPV calculation assigns a weighting function of i% to the money component. If we are interested in optimizing the combination of time and money, without skewing that optimization with a weighting function, then flush would appear to be more appropriate calculation.
96-04-22 jbg@CERF.NET (James Gambrell)
Tony - Your work in progress is interesting, but I beleive that you fail to consider some serious & coinsiderable issues.The outline that you have appears to attempt to convey a value of an investment for a non-public market. I do agree that the stock market appears illogical, but it is very rationale and very accurate. The value, at any given time, of a share of stock in the market is based upon an investor's perception of value (here investor is the leading or marginal investor that is a professional in the game.) That perception of value is made up of the expectations of free cash flows. This is comprised to two parts, current free cash flows (not to be confused with "earnings") and the expectations of future free cash flows. This will help explain why some companies trade at extremely high P/E multiples (by the way P/E multiples can be a poor barometer of value), as the investors believe that future free cash flows are to be great, then the current returns (dividends or current "earnings") are low and the emphasis is placed on times yet to come.
Yes the market seems illogical, but the lack of logic does not rest in the market, but perhaps with certain leading edge investors. People like you and me are merely price takers in the market - yet there are those marginal investors that drive prices one way or the other due to arbitrage opportunties that are instantly recognized and exploited down to zero.
Some of the particulars in your writing also need to be addressed. I do agree that what we have to invest is time and money. But how we invest is with information. While there is a near perfect flow of information in the market, the lag time for low-brow investors like us is deadly. By the time we learn something it has already been factored into the market. In fact, there is a lot of information factored into the market that we are not aware of. Over time, perhaps the best way to beat the market it to invest in the market basket,like a S&P500 index fund. If you want a higher return, then borrow funds and leverage yourself up to the amount of risk that you feel comfortable with. Yes there are certain funds that outperform the market, but out of 2,000 + mutual funds - which one is it - and will it be this year, next year or last year????
Your insight into believing the investment yielding the fastest time to "zero flush" is not correct. Your zero flush analogy is another way of describing the "payback" method of investment analysis. This is perhaps the worst method that there is. All methods have their shortcomings, as well as their proper place to be used. It is okay to utilize the payback method as merely one of many views of an investment. It does, however, ignore all investment returns garnered after "zero flush" is reached.
I would be more than happy to share with you the myriad of other popular methods used to evaluate investments.
I do believe that TOC can play a role in creating "expectations" of future company performance. Obviously, a company that uses TOC and the thinking processes will, at least in my opinion, perform better in the long run, but for this really to be of value to creating market wide expectations, the information must be communicated to the marginal investor, after all - they set the price or value in the company's stock on a daily baisis.
I always say, be careful what you ask for - you just might get it. You asked for opinions and feedback................
96-04-23 KLittle314@aol.com
Tim Sellan's response about definitions on how to measure an investments mentioned "other ways of calculating the time value of money." There's a fairly recent summary of an alternative to NPV approach, based on options models. Book is Investment Under Uncertainty by Dixit and Pindyck, Princeton U. Press, 1994. They explicitly account for "irreversibility, uncertainty, and the choice of timing." I just picked this book up again after a false start a year ago. It's for those of you who like books that assume you haven't forgotten calculus.
96-04-28 jbg@CERF.NET (James Gambrell) Subject: Re: Work in progress. (Rizzo)
Marginal Investors are the most sophisticated and informed of all investor classes. The ones operating at the "margin" or leading edge of the investment spectrum. These are the ones with the supercomputers, and other methods of sophisticated evaluation tools at their disposal.Most major securities, such as stocks and bonds, have funds and others that "follow" not only their particular industries but also the individual companies, so that when trends develop or news happens that effects or appears to effect the security's value they understand that and immediately take action to profit from that effect. Unfortunately, when guys like you and me read about something in the Wall Street Journal -- it is too late for it to do us any good. The effects are already factored into the market price.
It make perfect sense if you think about it. Suppose that you follow a particular company's stock very closely and you know that there is a good chance that something is going to happen (lets say a change in technology or regulation or even a new sales contract) that will effect the company's prospects. We know that the value of a company is based upon expectations of future free cash flows, it the expectation for these cash flows increases so will the value of the company. Obviously with this information, you would bid up the stock price - right up to the value of that future expectation of value. You would do this so as not to leave any money on the table. If you thought the stock was going to be worth $50, then you would be willing to pay all the way up to that $50 to get it. You wouldn't stop buying at $45, because you would be forging profits.
This is what happens every day, all day by thousands of marginal investors and funds. any opportune time to make these expected markets are almost instantaneously sucked up by the people that closely monitor the securities and the companies that issue them.
Alas, we are relegated to picking up the scraps, I'll be glad to elaborate more if you want, just let me know.
96-05-04 Caspari0@aol.com Subj: --Payback period
Your zero flush analogy is another way of describing the "payback" method of investment analysis. This is perhaps the worst method that there is. All methods have their shortcomings, as well as their proper place to be used. It is okay to utilize the payback method as merely one of many views of an investment.
Perhaps we are being too hard on the payback method. The calculation of payback is quick, easy, and understandable. Additionally, the reciprocal of the payback period approximates the internal rate of return (or DCF) of a project under certain conditions (i.e., high rate of return and economic life greater than twice the payback period).We do not often find the combination of these two characteristics in the Cost World (I have not seen a single instance of an example in an accounting or finance textbook illustration or exercise which has a rate of return greater than 30%). However, I routinely see capital expenditure opportunities with rates of return in excess of 150% in organizations which are implementing TOC. These are conspicuously desirable actions and it does not require "rocket science" to analyze them.
Why do we find these magnificant investment opportunities in TOC companies and not elsewhere? Simple. These organizations have learned to identify and focus on "Region III" actions (weak links of their chains) -- those areas where order of magnitude improvement is possible.
97-03-03 lleach@srv.net (Larry Leach) Subj: Flush & Project Management
David Kaser had the following comments from an off-line conversation. Thought more might be interested in the Flush description.> They have dropped Flush from the two day.
I would agree. However, maybe it is because I don't fully understand it.
[...]
I have finally become a Flush devotee. It took a long time, and was difficult for Dee Jacob, since I kept asking questions, and I think managed to confuse others in the class as well. I new writeup from Dr. Goldratt (I can't say where) made it suddenly clear, and I think I came up with an even better way to describe it. It is really very simple. Thus, another mark for Goldratt in elegance.
Think of a see-saw, with two people of unequal weights. The heavier one has to slide toward the fulcrum to balance. In Mechanical Engineering, we are balancing torque to zero at the fulcrum. The weights times their distance from the fulcrum are equal. Or, if you prefer, we are making the center of gravity move to the fulcrum to balance the see-saw.
With Flush, we want to value cash in hand, each day we have it (or don't have it.)
Flush is the same thing as the torque of the see-saw, except now the weight is dollars, and the distance from the fulcrum is days. (Makes a nice little figure.) Two cases with equal dollar-days balance. Its better to have more dollar-days (dollars are negative if outgoing); the see-saw tips in your favor. Thus, the big pile of gold close to the fulcrum balances the little pile all the way at the end if the Flush is equal. Makes total sense this way to me.
It may have been obfuscated through the time calculation. First understand the above, and the time calculation is simple.
97-03-03 lleach@srv.net (Larry Leach)
Attached please find the Excel 5 Macintosh file containing Flush calculations for the two project options. [ed note: I’m trying to figure out how to get this onto the web site – stay tuned. The two project options refer to the projects suggested by Tony Rizzo in the discussion at http://users.aol.com/caspari/l88.htm ]I had to make assumptions about investment and the time use of the bottleneck resource.
For those who can't [open the file], I assumed a $.5m/month investment in each project (i.e., $9 m total). I assumed that the bottleneck resource causes a start delay of the other activity of the activity time in the first project. B first is a little better, but not much.
Up the project investment to $18 m ($1 m/month), and both projects are losers; they never flush.
I didn't try to find the investment limit.
97-03-04 @lucent.com (Anthony R Rizzo)
With Flush, we want to value cash in hand, each day we have it (or don't have it.)Most people don't have much of a problem with calculating flush. But they do have a big problem with their interpretation of flush. So, here's a question that's been on my mind for some time.
When a project is said to reach the zero-flush point, exactly what have we accomplished?
97-03-08 jeaster@verinet.com (James Easter)
I'd appreciate a discussion how flush is fundamentally different from Net Present Value-based metrics-I know they are different in formula-but how would I make a different decision using one vs. the other-on any given project or potential investment proposal?I would especially appreciate an example where using NPV-metrics would lead me to the wrong decision and flush would prevent it-or an example of where flush would lead me to do a project/proposal and NPV would not.
97-03-08 Caspari@aol.com
I'm not quite sure how I would evaluate whether a decision were right or wrong with NPV or Flush -- the "right" answer seems to depend on the criterion used to evaluate the "rightness" (i.e., Flush or NPV).I have not seen a complete description of FLUSH in the literature, but hope that Goldratt's promised new book will explain it further. [...] The FLUSH measure appears to rely on the calculation of dollar*days in a manner consistent with the other "dollar*day" metrics (i.e., inventory dollar days (IDD) and throughput dollar day (TDD)) included in the basic TOC measurements.
The IDD and TDD metrics appear to me to be useful in the area for which they are prescribed (see Chapter 24 in The Haystack Syndrome). The FLUSH use of dollar days appears to be somewhat different than the IDD and TDD use in that the IDD and TDD metrics are used to reveal patterns, while the FLUSH measurement appears to be used in terms of its value. A negative value of dollar*days indicates an undesirable project, and given a choice between two projects the one with the larger positive dollar*days is preferable. The "dollar*day" metrics have bothered me a little because they appear to give equal weighting to the monetary measure and to the unit of time. For example, ($2 * 3 days = 6 dollar days) and ($3 * 2 days = 6 dollar days). The argument for dollar*day measures that I have seen is that both time and money need to be considered. But, I have not seen any argument presented which supports an equivalency between the two.
It seems to me that this may be a weakness in the Flush metric. At any rate, it is different than the more conventional assumptions used with discounted cash flow analysis (i.e., both time and money are important, but that there is a cost of money--known as the cost of capital, discount rate, or interest rate--which is different in different circumstances).
As to the question of an example, consider the following investment proposal:
Initial investment at present time = $50,000
Cash inflow received at end of first year = $30,000
Cash inflow received at end of second year = $28,000
Cash inflow received at end of third year = $26,000How would we evaluate this?
Using Net Present Value: If the cost of capital (discount rate) is 5 percent, then the proposal is desirable from a NPV point of view (i.e., NPV = $26,428, which is positive). NPV says make the investment in this case.
Using FLUSH: At the end of the third year, the outstanding dollar*days are a negative 22,644,880. Since the dollar*days do not reach zero (I think that zero is the "flush" point), the FLUSH metric, I assume, tells us not to make the investment.
Clearly, at least one of the techniques is fatally flawed.
I, too, would like to see more discussion of this FLUSH metric.
97-03-08 tocguy@lucent.com (Anthony R Rizzo)
>I would especially appreciate an example where using NPV-metrics would lead me to the wrong decision and flush would prevent itIt is most unfortunate for me, that I can satisfy your request with a personal example. Had I understood flush at the time that I made the decision described below, I assure you that my decision would have been very different. Nevertheless, here's the example.
I sold a piece of property some years ago. The sale price was $170K. The buyer asked me to carry a mortgage for $125K, which I did. At that time, the going interest rate was 10%. I agreed to carry a ten-year mortgage at 10%, for $125K.
Shortly after, I nearly got laid off. One thing that I'd contemplated for some time was to start a business. But to start the kind of engineering business consistent with my capabilities at that time, I would have needed a fair amount of cash on hand, and I would have needed it fast. I have to report that the monthly mortgage checks, which the NPV metrics say is equivalent to the present value of the $125K, did not provide me with the opportunity to start that business. The Flush measurement would have steered me to a different decision, regarding the mortage.
dsirias@cnu.edu (Danilo Sirias)
Before we go into further discussion, can somebody explain how flush is calculated?
97-03-08 tocguy@lucent.com (Anthony R Rizzo)
Somebody asked how to calculate flush:--- ---
| |
| /later |
| | |
limit | | (throughput_rate x time) d_time |
as later | | |
goes to | /now |
infinity | |
--- ---
97-03-09 fps_toc@juno.com (Frank P Sansone)
For those of us not very familiar with the FLUSH calculation, could you use the example of the 10 year mortgage and show us how the FLUSH concept would lead to a different conclusion? I think it would be very helpful to me, and perhaps some others on the list.
97-03-09 tocguy@lucent.com (Anthony R Rizzo)
After John Caspari's last message, I've had to ask myself a question or two regarding the flush calculation. I've also had a recent discussion with another TOC practitioner, recently, who's been trying to help me understand the concept better. I think that there are a number of underlying assumptions, accompanying the current calculation, which have not been discussed at all. I'll try to bring them to light. Who knows? This may be a learning experience for me too.I'm going to approach this from the point of view of myself, an investor with very limited amounts of two resources, money and time. As an investor, in fact, these are strategic resources, in the sense that I can't do very much to elevate their capacities. The limited capacity of the money resource is something that we all encounter, almost daily. But the limited capacity of the time resource is one that we often treat as being less important. In fact, I'm 44 year of age now. I won't be on this world for very long. So, time is also important to me. I want to use it as efficiently as I use my money resource.
By the current flush calculation, the flush measurement for my mortgage example would go like this.
Let's say that the investment is x dollars on day 1.
Let's say that each month, for 120 months, I get f(x,120,i) dollars from the mortgage, where f(x,120,i) is the monthly amortization payment, and i is the interest rate.
At the end of the 1st month, the flush measurement is:
-x * 1 in units of dollar-months.
At the end of the 2nd month, the flush measurement is:
(-x * 2) + f(x,120,i) * 1
At the end of the 3rd month, the flush measurement is:
(-x * 3) + f(x,120,i) * 2 + f(x,120,i) * 1
At the end of the nth month, the flush measurement is:
(-x*n) + f(x,120,i)*(n-1) + f(x,120,i)*(n-2) + ... + f(x,120,i)*(1)
where n <= 120.
But, at this time I have a problem with using this calculation, because I'm not certain about all those underlying assumptions. It's like using a mathematical model of a physical system, without knowing all the limitations of the model. In engineering, that's always a dangerous and foolish thing to do.
As an investor with limited capacities of the money and time resources, I want to use both of these most efficiently. Efficient use of the money resource is something to which we've all been exposed. Typically, we go for the investment with the greatest rate of return. But what about the efficient use of the time resource?
So, again, let's take my case. I expect to work until age 64. At that time, I would like to stop working and begin enjoying my waking hours (wow! what a thought. Imagine if we could enjoy our sleeping hours). How can I ensure that my investment decisions make optimal use of the COMBINATION of my two limited resources?
A few weeks ago, on this very list, there was a discussion about the optimization of something or else. One person pointed out that the conventional approach to the optimization of two quantities simultaneously was really a compromise. Typically, we multiply the two quantities, and we optimize the product. Then, we can be sure that we haven't shortchanged either of the two quantities. But if we multiply money and time, what do we really have? In other words, what's the significance of this product?
In geometric terms, of course, it's just an area. This is somewhat meaningless. But in terms of my two limited resources, the product can be interpreted as a measure of how much of the combination of time and money I have. The product can be interpreted as the capacity of my time*money resource. At the time that I was contemplating carrying the mortgage, the capacity of my time*money resource was ($125K)*(20years).
Yes, the life of the mortgage is 10 years. But the time that I have available for using my two resources is 20 years (recall, I'll retire at age 64, and I'm 44 now). I need to know how efficiently I am using my entire combined resource, don't I? Therefore, I really should extend all calculations, aimed at comparing multiple investments, out to 20 years. The far end of that 20 year interval, in other words, is my fixed reference frame. I need to know (or estimate) the capacity of my combined money*time resources at the end of that 20th year.
In practice, this is probably not necessary in the majority of cases. Most often, we need to calculate only the time needed to achieve a complete recovery of the money*time resource that competing projects require. Most often, the project that provides us with this complete recovery of our limited combined resource in the shortest interval greatly outpaces the competing projects for the remainder of the 20 years, or for whatever number of years represent the extent of the time resource.
But there are exceptions. Certainly, we can think of exceptions. My pen pal in California, for example, asked this question, "If project A reached its zero-flush point in three months, and project B returned $1 Billion a day later, flush would have me choose project A."
Understandably, he has a problem with this. He should! The use of the zero-flush point alone implies that the project that reaches the zero-flush point first outpaces the other candidates for the remainder of the investor's available interval. This is an assumption inherent in the use of the zero-flush point as a comparison tool. As is true in all instances when we make assumptions, we need to vefiry them.
However, if my pen pal thought in terms of the capacity of the money*time resource available to him some time in the future, I'm certain that he would conclude that the project that returns $1 Billion one day later is the project of choice. For this contrived example, the assumption inherent in the use of the zero-flush point doesn't hold. Therefore, why would we make this assumption?
Gee, I hope that this helps. I've reached the end of my explanation resource. This is not to say that I'm being impatient. I'm not! I just don't know how else I can explain it.
97-03-09 elyakim@NETVISION.NET.IL (Eli S)
I want to start with the example John Caspari had used:Initial investment: $50,000.
Cash inflow at end of first year: $30,000
Cash inflow at end of second year: $28,000
Cash inflow at end of third year: $26,000John has calculated that at the end of third year the outstanding dollar*days are: -22,644,880. He concluded that the Flush analysis would not recommend this investment. I fully rely on John's calculation. But, Flush analysis will only say: every day after the end of the third year will add $34,000 (30000+28000+26000-50000) to the balance. So, the return time for the dollar-days investment is less than 5 years. Is it good? Maybe, I don't know what are the alternative investments for the $50,000. Something similar for Tony's example: after 120 months you still add the fixed sum that is your actual cash flow balance every day - until the dollar-days balance becomes 0. Or maybe it has reached 0 dollar-days before the 120 months have passed. Again, is it good or bad?
I prefer the NPV or the Internal Rate of Return upon the Flush measurement. At least I understand the logic behind the two. The notion of "cost of money" makes full sense for me. When money is considered as a constraint we want to exploit it. We know we can safely get the regular interest rate - so why not compare the two alternatives?
The criticism on the NPV is its dependency on the interest rate used. I think this is realistic - because the value of the invested money is really depended on how much the invested sum is worth for you to give it away for a period of time. This is the meaning of the "cost of money".
As John has pointed out the assumption, of the Flush metric, that time and money are equal is a very strong one. The other strong and even arbitrary assumption is that the time of full dollar-days return is of special significance in comparing alternative investments. This assumption means that any incoming cash that arrive after the dollar-days have been fully returned is irrelevant for the decision.
Investments are difficult to decide upon not because we do not understand the difference between Flush and NPV or IRR. Investments are difficult because it is never certain. Who really knows what will be the inflow cash? Both quantities and exact timings are usually very big "maybe" or "hopefully". This is, to my view, the real issue: how to make a sound decision when everything is uncertain. The statistical analysis needs data that we usually don't have. Statistics is also helpless when there is a lot of partial dependencies between the variables - as our reality usually is.It is time to develop a probability process that will be based on our intuition - because in the vast majority of the cases that's all we have. For me, Flush or NPV, are "chupchics" (meaning - small matter, not important).
97-03-09 tocguy@lucent.com (Anthony R Rizzo)
As John has pointed out the assumption, of the Flush metric, that time and money are equal is a very strong one.Indeed, it is. Is it unjustafiably strong?
The other strong and even arbitrary assumption is that the time of full dollar-days return is of special significance in comparing alternative investments.
(165) A bottleneck is any resources the capacity of which is less then the demand placed upon it.
If (165) A bottleneck is any resources the capacity of which is less then the demand placed upon it,
and if (167) a bottleneck prevents more progress toward the goal of a system,
then (168) the bottleneck is a critical constraint resource for the system.(170) Most investors have no resources other than money and time, with which to invest in projects.
If (170) Most investors have no resources other than money and time, with which to invest in projects,
then (180) for most investors, money and time are their only resources.(190) Most investors have available to them more investment projects than they have money or time with which to invest in the projects.
If (190) Most investors have available to them more investment projects than they have money or time with which to invest in the projects,
and if (180) for most investors, money and time are their only resources,
and if (185) projects place a demand on the money and time resources of investors,
then (200) for most investors, money and time are bottleneck resources.If (210) the goal of most investors is to make as much money as possible within a finite period of time,
and if (200) for most investors, money and time are bottleneck resources,
and if (220) for most investors, money and time prevent further progress toward the goal of the investors,
then (230) for most investors, money and time are critical constraint resources.(100) The money*time calculation is a measure of the capacity of the combination of the money and time resources available to an investor.
If (100) The money*time calculation is a measure of the capacity of the combination of the money and time resources available to an investor,
and if (230) for most investors, money and time are critical constraint resources,
then (105) the money*time calculation is a measure of the capacity of an investor's critical constraint resources.If (105) The money*time calculation is a measure of the capacity of an investor's critical constraint resources,
and if (110) the interval that a selected investment requires, before it restores an investor's critical constraint resources to the same level that the investor has at the time that the investor makes the investment, is shorter than the interval required by other potential investments,
then (120) the seleced investment releases the investor's critical constraint resources completely, before the other potential investments release the investor's critical constraint resources completely.If (190) Most investors have available to them more investment projects than they have money or time with which to invest in the projects,
and if (230) for most investors, money and time are critical constraint resources,
then (130) upon the complete release of his/her critical constraint resources the investor is able to employ his/her critical constraint resorces again, with additional investments.If (130) upon the complete release of his/her critical constraint resources the investor is able to employ his/her critical constraint resorces again, with additional investments,
and if (120) the selected investment releases the investor's critical constraint resources completely, before the other potential investments release the investor's critical constraint resources completely,
then (140) the first investment lets the investor exploit his/her critical constraint resources more completely than do the other potential investments.This assumption means that any incoming cash that arrive after the dollar-days have been fully returned is irrelevant for the decision.
If (150) an investment continues to return cash to the investor, after it releases the invetor's critical constraint resources completely,
then (160) the investment increases the capacity of (elevates) the investor's critical constraint resource.(300) The zero-flush interval is the interval required for an investment to completely releases an investor's critical constraint resources.
If (300) The zero-flush interval is the interval required for an investment to completely releases an investor's critical constraint resources,
and if (310) for some investments the zero-flush interval is infinite,
then (320) some investments never completely release the investor's critical constraint resources.If (330) some investments never completely release the investor's critical constraint resources,
then (340) some investments waste forever a portion of the investor's critical constraint resources.
97-03-10 fpatrick@eclipse.net (Francis S. (Frank) Patrick)
John Caspari wrote:The argument for dollar*day measures that I have seen is that both time and money need to be considered. But, I have not seen any argument presented which supports an equivalency between the two.
A subsequent message from Tony Rizzo included:
(100) The money*time calculation is a measure of the capacity of the combination of the money and time resources available to an investor,
and if (230) for most investors, money and time are critical constraint resources,
then (105) the money*time calculation is a measure of the capacity of an investor's critical constraint resources.When we use the 5 focusing steps, we sometimes have to address the need/possibility to ELEVATE the constraint. If the combination of money and time is the constraint preventing more throughput of the system, then in order to elevate the constraint, we need to acquire more of either time or money to apply to the endeavor. Acquisition of time while maintaining a target completion date could most easily be interpreted as buying more resources to apply to the effort...which requires more money.
And actually, the acquisition of money to elevate the constraint in itself is useless unless that money can be productively applied to the needs of the project, i.e., buy the time/capability to complete the project. So the old maxim "time is money," while appropriate for thinking about making money, could be switched to "money is time" when thinking about meeting due dates.
While this line of thought suggests an "equivalency between the two," I'm not completely convinced myself. Since the purpose of money in this case is to buy time, I would actually suggest that of the two, time takes primacy as a constraint, since it can, in many circumstances, be easier to find additional funding (you could always be willing to pay a high enough price (interest) to your local loanshark) that to acquire and/or fully utilize additional time.
But...There is one additional point I'd like to bring up. Both of the methods, traditional NPV or non-traditional Flush, depend on one aspect of the project/proposal which, at least in my experience, is the bane of all such decisions...the validity of the numbers going into the calculation. How much will actually be spent on a project over a course of a couple years? How much will we make from the new endeavor once complete? These are projections, and even in a contract situation, you can predict only one of them with even a modicum of accuracy.
(I remember a situation at an old consumer products employer of mine where the engineering staff was being run ragged doing ridiculously accurate early estimates of projects while a swing in even a fraction of the market projections would dwarf an even 20-50% variation on cost.)
Neither of the calculations can effectively be used without taking the variability of spending/throughput into account, so we end up comparing a range of possible outcomes from one decision to a range of possible outcomes to an alternative decision. And in the potentially significant intersection of these outcomes, the decision comes down to which set of numbers you have more faith in and which side of the coin comes up after the flip.
97-03-10 jlevy@mdc.com (Joe Levy)
The example of flush that Tony Rizzo gave about the taking the mortgage versus cash made me think that Flush is a simplified view of the world that gets us thinking in terms of the right decision much as T,I, and OE can. It is not a really sharp pencil however. If Tony had really wanted to start his business he could have borrowed the money using his paper on the house as collateral, because bankers use NPV. However if interest rates had gone up or down in the meantime Tony would have either more or less money as a result. This was part of the risk he assumed with the paper, but he sold the house and turned a drain into a monthly asset. John Caspari's 50K investment returning 84K in three years is a 19% annual return on average; perhaps a little small in many businesses but ok in some. Flush it seems gives weight to cash flow along with return on investment (ROI). Its quick, its dirty, I think, however, that a business decision must be based, in the final analysis, on ROI, with the constraint that the cash flow is affordable (that is you don't bankrupt wanting for return or the cost of money doesn't cut the true return too much). I seem to be back at T, I, and OE.
97-03-10 elyakim@netvision.net.il (Eli S)
It seems a nice debate is developing between myself and Tony and Larry. While, being an Israeli, I like to argue, I also want to emphasis again that the whole question whether to use Flush or NPV or IRR is not too important anyway. we know too little about the actual outcomes of an investment/project.Tony argues that both money and time are constraints. Suppose it is. When we speak about our time being a constraint -- this is NOT dealt by the Flush. The one constraint we deal with Flush is money. Tony is right that the money * time is a measure for capacity. We do the same with resource constraint: we load it according to the constraint's time. Time is a factor to be associated with any constraint. The way we exploit our time is a different thing - even though lack of money (interactive constraint) may be problematic. When you invest 1M dollars and get it back after a year --- when did you recover your capacity? It seems to me that it is fully recovered after 1 year - from capacity point of view. You haven't got any T to justify your investment, yet. But, you have your capacity for use from that time onward.
Suppose a 1M investment and 1 payback of 2M after a year. When the payback comes in - the capacity is back and even doubled. The dollar-days balance is zeroed only after two years. But, nothing really happens after two years - certainly one has restored the full capacity long time ago. There is a difference between restoring capacity and generating the full value for the investment in capacity.
If money is to be treated as a constraint: I think that in order to make sound decisions I need to know the limit imposed on me. Flush doesn't consider it at all. >From dollar-days point of view 1M for a year is exactly the same amount of DD as 0.5M for two years (John Caspari's point). Suppose all you have is 1M (actually, for me this is not that bad). Considering all the other needs - there is a difference between the two investments. 0.5M seems less risky and less limiting my ability to make business. Neither Flush nor NPV take it into account. Stating the interest rate in the NPV equations may include one's assessment for risk. Not a very structured algorithm - but, at least one can put in some judgement. The way I was taught about NPV is that the decision on what interest rate to use is a complex one - and certainly not simply the current profitability of the company, where it is not certain that an additional investment in the company will yield the same profitability (and what about a company that right now loses money?).
Larry, I don't follow your argument that Flush and NPV are different things. BOTH ARE USED TO SET PRIORITIES TO ALTERNATIVE INVESTMENTS! I don't know of no other decision that I need either. I don't know why you state that NPV doesn't relate to cash flow. Both, Flush and NPV, are using the same data input about investments and payments - quantities and timings. Neither of them take shortage of cash into account.
I also don't see the relevancy of the torque example. Do I look for equilibrium here? And what does that equilibrium mean?
Let's take a simple example:
1M investment at the beginning of year 1.
1M payback at the beginning of year 2.
1M payback at the beginning of year 3.That's investment-1.
Now investment-2 is the same as investment-1 with one additional payment at the beginning of year 4.
To me investment-2 is by far preferable. What does Flush say about it?
At the beginning of year 2: the dollar-year metric is 1M-year. This balance is kept the same throughout the second year. After the third year payment - the dollar-years goes down. At the beginning of the fourth year the dollar-years is zero. All in all three years return of dollar-days (dollar-years). That means that both alternatives are the same from the Flush judgment.
My main concern with Flush is that it causes a lot of confusion. This is by far the most devastating negative branch of the idea.
Also, please don't interpret the above as if I'm opposing the idea of the dollar-days more generic concept. I think the measurement of Thoughput-Dollar-Days can be quite effective in pointing to deviations from a plan. That idea can be expanded to other areas as well.
97-03-10 Caspari@aol.com
On 97-03-09 elyakim@NETVISION.NET.IL (Eli S) wrote:John [Caspari] has calculated that at the end of third year the outstanding dollar*days are: -22,644,880. He concluded that the Flush analysis would not recommend this investment.
I fully rely on John's calculation. But, Flush analysis will only say: every day after the end of the third year will add $34,000 (30000+28000+26000-50000) to the balance. So, the return time for the dollar-days investment is less than 5 years. Is it good? Maybe, I don't know what are the alternative investments for the $50,000.
I apologize for using the wrong decision criterion for FLUSH. Eli is correct - the zero flush point is 4 years and 301 days. Again, I will take a guess at the decision criterion associated with FLUSH and suggest that perhaps we should compare the investment with a cost of capital rate of return of 5 percent (the alternative use of capital). In this case the cash flows would be:
Initial investment: $50,000
Received at end of first year: $2,500
Received at end of second year: $2,500
Received at end of third year: $52,500A second way of looking at this would be to have a perpetual annual cash inflow of $2,500.
The FLUSH point in the first case is at 22 years, and in the second case at 41 years. Using the criterion of the earlier FLUSH point, FLUSH would give the same signal as NPV. Thus my original criticism does not hold based on this example.
I prefer the NPV or the Internal Rate of Return upon the Flush measurement. At least I understand the logic behind the two. The notion of "cost of money" makes full sense for me. When money is considered as a constraint we want to exploit it. We know we can safely get the regular interest rate - so why not compare the two alternatives?
I do not see the advantage of FLUSH over NPV or the Internal Rate of Return (IRR) at this point. However, knowing the depth of my Cost World roots, I would like to understand why some people think FLUSH to be far superior. I agree with Eli S that the whole issue is probably a chubchik in most cases. However, if there should happen to be cases in which it is not a chubchik, I would like to know what the characteristics are.
The first question I have is the same as James Easter raised originally: Does FLUSH give different signals than discounted cash flow methods. So - let me try another example.
Assume that we have two proposals before us. They have the following cash flow characteristics:
Proposal A:
Initial Investment: $50,000
Received at the end of each of years 1 - 14: $5,000
Received at the end of year 15: 55,000Proposal B:
Initial Investment: $50,000
Received at the end of year 15: $208,862.41Each of these proposals has an IRR of 10 percent. Thus in this sense they are equivalent.
However, Proposal A reaches a FLUSH zero point at 18 years and Proposal B reaches a FLUSH zero point at 19 years and 263 days. I assume, then, that FLUSH prefers Proposal A.
Now, let us increase the return of Proposal B to about 11 percent (actually 10.94515%) by adding an additional $28,600 to make the payment at the end of year 15 be $237,462.41. The FLUSH zero point is now 19 years. FLUSH still prefers proposal A.
It is interesting to note, however, that Proposal A adds $75,000 per day to the dollar*days after the fifteenth year and Proposal B adds $187,462.41 per day to the dollar*days after the fifteenth year. The effect of this is that after the 20th year Proposal B has a larger amount of positive dollar*days than Proposal A. For example, by the 30th year Proposal A accumulates 328,716,000 dollar*days while Proposal B accumulates 753,101,551 dollar*days.
I have some difficulty interpreting the FLUSH metric. It seems that although FLUSH considers both money and time, it does not take into account the time value of money. If the time value of money concept is as erroneous as the product cost concept, then I am in for another MAJOR paradigm shift!
13 Mar 1997: tocguy@lucent.com (Anthony R Rizzo)
Eli Schragenheim writes:It seems a nice debate is developing between myself and Tony and Larry. While, being an Israeli, I like to argue, I also want to emphasis again that the whole question whether to use Flush or NPV or IRR is not too important anyway. we know too little about the actual outcomes of an investment/project.
Argue? Nah! It's evident that we have different opinions regarding one particular subject. I'd simply like to discuss this issue, so that we can arrive at a common understanding. Like John Caspari, if I'm missing something, then I want to know what it is. So, let's see if we can't achieve a clearer understanding of this thing.
So far as knowing too little about the actual outcomes of an investment project is concerned, I agree. It's very tough to predict the future. But is it really always necessary to be able to predict the future, when one is choosing investment options?
Tony argues that both money and time are constraints. Suppose they are.When we speak about our time being a constraint -- this is NOT dealt by the Flush. The one constraint we deal with Flush is money.
I have a clarity reservation regarding this statement. Why do you say that time is not addressed by the flush calculation? I don't understand this.
Tony is right that the money * time is a measure for capacity. We do the same with resource constraint: we load it according to the constraint's time. Time is a factor to be associated with any constraint. The way we exploit our time is a different thing - even though lack of money (interactive constraint) may be problematic. When you invest 1M dollars and get it back after a year --- when did you recover your capacity? It >seems to me that it is fully recovered after 1 year - from capacity point of view. You haven't got any T to justify your investment, yet. But, you have your capacity for use from that time onward.
I have an entity existence reservation with the conclusion that recovering $1M after a year restores the investor's original capacity.
If flush is a measure of the capacity of an investor's money*time constraint,
and if after 1 year the investor recovers the original $1M from the investment,
then the measure of the capacity of the investor's money*time constraint, at the time that the investor recovers the $1M, is 365 million dollar-days smaller than the measure of the capacity of the investor's money*time constraint at the time that the investor makes the original investment.If the measure of the capacity of the investor's money*time constraint, at the time that the investor recovers the $1M, is 365 million dollar-days smaller than the measure of the capacity of the investor's money*time constraint at the time that the investor makes the original investment,
then the recovery of the $1M one year later does not restore the capacity of the investor's money*time constraint to the level that the investor has at the time that the original investment is made.Suppose a 1M investment and 1 payback of 2M after a year. When the payback comes in - the capacity is back and even doubled. The dollar-days balance is zeroed only after two years. But, nothing really happens after two years - certainly one has restored the full capacity long time ago. There is a difference between restoring capacity and generating the full value for the investment in capacity.
I think that much of the confusion is caused by our lack of an easily understandable reference frame for the time component. The current flush calculation requires that we determine the changes in flush from the time that the investment is made. Mathematically, this is not a problem, because we are calculating invariant quantities. But conceptually, this can be very difficult to digest.
For me, it's much easier to think in terms of the capacity of the money*time constraint available to the investor at any one instant. To be able to calculate this, we need some instant in the future as a fixed reference frame.
The magnitude of the actual interval is irrelevant, so long as the interval is sufficiently long as to extend beyond the period when the investments are expected to return cash flows. So, let's use 10 years as the interval available to the investor, just for the sake of this discussion. Let's say that the investor has available $1M and 10 years, at time zero (boy, that does begin to sound like a capacity, doesn't it?).
At time zero: F = $1M * 10Y = 10 M$Y is the capacity of the constraint.
At time 1 yr: F = ($1M * 10Y) - ($1M * 1Y) + ($2M * 9Y) = 10 - 1 + 18 = 27 M$Y
It would seem that this investment greatly elevates the capacity of the investor's money*time constraint. Also, note that the capacity of the constraint decreases steadily with time, if the investor does nothing with the money.
If money is to be treated as a constraint: I think that in order to make sound decisions I need to know the limit imposed on me. Flush doesn't consider it at all.
I need clarity. What limit? On what is this limit imposed?
From dollar-days point of view 1M for a year is exactly the same amount of DD as 0.5M for two years (John Caspari's point). Suppose all you have is 1M (actually, for me this is not that bad). Considering all the other needs - there is a difference between the two investments. 0.5M seems less risky and less limiting my ability to make business. Neither Flush nor NPV take it into account. Stating the interest rate in the NPV equations may include one's assessment for risk. Not a very structured algorithm - but, at least one can put in some judgement. The way I was taught about NPV is that the decision on what interest rate to use is a complex one - and certainly not simply the current profitability of the company, where it is not certain that an additional investment in the company will yield the same profitability (and what about a company that right now loses money?).
I understand the limitations that you verbalize. But I don't see why I would use NPV instead of flush, given this last paragraph. Are you suggesting that NPV is the method of choice, because by adjusting the interest rate one can somehow account for risk? This is just a weighting factor, isn't it?
13 Mar 97 "Mark Woeppel" <Mark_Woeppel@msn.com>
Tony writes:>If (190) Most investors have available to them more investment projects than they have money or time with which to invest in the projects,
and if (180) for most investors, money and time are their only resources,
and if (185) projects place a demand on the money and time resources of investors,
then (200) for most investors, money and time are bottleneck resources.This speaks directly to John's assumption that money and time are of equal importance. Therefore, I have a clarity reservation on entity #180.
My two cents is that if we focus on the time=money issue, we miss the point. I think the only constraining resource is money. We take for granted that time is limited. Just as Throughput is a rate, return on investment is a function of the time the investment is held.
Flush gives a different slant on the return time of the investment. Therefore, its usefulness is not in comparing it to NPV or other types of analysis, but to other investments for which the flush calculation has been done.
13 Mar 1997 tocguy@lucent.com (Anthony R Rizzo)
My two cents is that if we focus on the time=money issue, we miss the point. I think the only constraining resource is money.OK. Let's see. I have $10K today, and I need $100K to meet a necessary condition. Am I more likely to meet my necessary condition if I have 1 day available or if I have 15 years available?
Is money really the only constraint resource?
We take for granted that time is limited.
Yes, we do. But does this mean that we really take time into account when evaluating investment alternatives?
Just as Throughput is a rate, return on investment is a function of the time the investment is held.
Yes, this is true.
Flush gives a different slant on the return time of the investment. Therefore, its usefulness is not in comparing it to NPV or other types of analysis, but to other investments for which the flush calculation has been done.
We agree!
13 Mar 1997 Peter Skelton <pgs@adan.kingston.net>
The constraint is the resource/facility whose limitation limits throughput.Time and money are two raw materails to an investor. Clearly in some situations time will be the constraint and in others money.
Suppose you had steel bar stock you have to run through a lathe. Clearly whether the bar stock or the lathe are the constraint depends on the amount of stock available, the amount of machining to be done, haw fast the lathe can run and other factors. Arguing about it in the abstract is pointless.
13 Mar 1997 lleach@srv.net (Larry Leach)
Tony Notes,I understand the limitations that you verbalize. But I don't see why I would use NPV instead of flush, given this last paragraph. Are you suggesting that NPV is the method of choice, because by adjusting the interest rate one can somehow account for risk? This is just a weighting factor, isn't it?
We'll let Eli answer for himself, but it is my observation (and I believe, one of the bases for the layers of resistance) that our minds are FILO (First In, Last Out) devices. That is, whatever we learned first has intrinsic value and does does not require justification. The new idea always has to defend itself against the old, not the other way around. This has nothing to do with the correctness of the idea. Just the way we are. I know I went through exactly the same denial we are hearing on Flush.
Its not that we don't want to learn or consider things on an equal basis. Its the way we are built. It has to do with the self-reinforcing nature of our belief systems.
Just ask Galileo. (Oh, he's not around any more. Died in Jail, didn't he?)
Our problem with this discussion is that we have failed to get through layer one with all participants. Therefore, working and discussion the other layers (while kind of fun) is doomed to be fruitless.
So, how about we start to see if there is a problem to start with, and one that is in our control?
1. We have learned, and presumably all agree on the TOC principles of T, I, and OE.
2. We are addressing the tool we need to make Project investment decisions.
3. The Project Decisions we are discussing are those involving time and money. (There are lots of others that involve other things.)
4. Dr. Goldratt has proposed requirements for such a tool. They are that the tool should respond positively when:
4.1 Less money is spent.
4.2 Less time is taken.
4.3 Cash is tied up for less time.
4.4 Throughput is gained sooner.
4.5 Money is spent later.
4.6 The real cost of money is less.Do we all agree on these requirements?
I will continue assuming agreement. Let me know if we are stuck here. If we are stuck here, we should not wate our time to go further.
I think we can also agree that there are many tools that satisfy 4.1 through 4.5, including our old friend NPV (regardless of what discount rate we use.)
So, we seem to be focused on item 4.6, "The real cost of money."
Here is where Tony's constraint argument comes in, and where we need to focus. Do we have agreement on this?
Now, let's follow Tony's idea, and say we have a pile of money. What is it? It certainly isn't T. It's not OE unless we use it for that. So it must be I.
When I put it into a Project, what happens? Doesn't it become OE?
Certainly if it is an internal product development. What if it is a contract project we are doing for someone else?
So I goes down (I guess we like this in general, but it is hard to understand when it is cash!), and OE goes up (we don't like this). There is no immediate impact on T. What is the 'real' cost of this money?
Don't see any 'interest' involved in the calculation at all, so it id difficult to see how interest or interest rates should have anything to do with it.
Net profit went down by the amount we invested, by adding to OE. (NP=T-OE) It goes down on the day we spend the money. It does not come back to us until we get the expected increase in T. So we are sitting with the reduction in NP until that happens.
Because of requirement 4.3, it must matter if that T happens tomorrow, or six months from now. How do we understant this difference?
What is the 'real' cost, if I get it back six months from now, vs. getting it back tomorrow?
(Keep in mind that we are talking a Project decision making tool, that is looking at alternatives of some kind. It might be to do the project or not. It might be to spend more to accelerate the project. It might be to choose between project alternatives. It might be to invest more or less, and therefore influence our T expectations; in amount or timing.)
Let me think about it. Open for more ideas!
Mark Woeppel noted:
My two cents is that if we focus on the time=money issue, we miss the point. I think the only constraining resource is money. We take for granted that time is limited. Just as Throughput is a rate, return on investment is a function of the time the investment is held.
Well, what if the project is to develop the 'P9' clone chip. Is it worth anything to have it out six months before Intel? Might it affect market share, and the price we can charge? (How much is a 486 chip today, anyhow?)
14 Mar 97: "Mark Woeppel" <Mark_Woeppel@msn.com>
Larry writes in response to me:>My two cents is that if we focus on the time=money issue, we miss the point. I think the only constraining resource is money. We take for granted that time is limited. Just as Throughput is a rate, return on investment is a function of the time the investment is held.
Well, what if the project is to develop the 'P9' clone chip. Is it worth anything to have it out six months before Intel? Might it affect market share, and the price we can charge? (How much is a 486 chip today, anyhow?)
Of course it would! But you're missing my point.
I'm thinking that time is ALWAYS the ultimate constraint. It cannot realistically be lengthened (as far as I know). Therefore, it need not be considered a constraint separately.
My point was that we are trying to choose the "best" solution among multiple options - regardless of how we come to that conclusion (what is "best"). Thus, an option that returns money sooner or at a higher rate over the long term is the "best" solution (check your goal assumption here:))
When we consider the case sited and the other one I saw (I forget), the product/market/process is immaterial. If we have a case where we reduce our time to market and in the process, throughput will increase sooner and/or greater over another scenario, then we will balance the time to return our investment with the greater throughput.
The "best" scenario is one that gets us more throughput in the least amount of time with the least amount of investment.
14 Mar 97 Eli S <elyakim@netvision.net.il>
As the discussion develops I need to react to Larry. I don't seem to remember that the categories of legitimate reservations contain the 5 layers of resistance. Used as an argument this is a knife with double blades. I could have claimed that as Larry was recently converted to the cause of Flush, he is under the spell of Cognitive Dissonance. I think it is unfair to use such arguments and it is also not practical.I'll stick to the rational arguments raised by Tony. I believe we can discuss the matter until we feel we understand the rational and the underlining assumptions.
In my previous post I wrote: "When we speak about our time being a constraint --- this is NOT dealt by the Flush. The one constraint we deal with Flush is money".
Tony had reacted:
I have a clarity reservation regarding this statement. Why do you say that time is not addressed by the flush calculation? I don't understand this.
I've tried to address a previous post by Tony, where he's described both money and time as OUR personal constraints as human being. How do you exploit your time is not addressed by Flush. How you treat your own money MAY be addressed by Flush.
Time in itself is not a constraint. It has to be linked to a constraint to be a capacity constraint. "We don't have time" implicitly says that a certain resource cannot produce what is required within the time frame.
Flush (and NPV as well) is a method to measure the exploitation of money as a constraint. As Mark also states that Flush and NPV cannot be directly compared - let me submit that BOTH ARE USED FOR THE SAME PURPOSE - to set the investments priorities when money is constrained. When two different methods are supposed to do the same: COMPARING THE TWO IS INEVITABLE.
In my previous post I've agreed to Tony's claim that the capacity of money is money*time. Thinking a little bit more had convinced me otherwise. Capacity is a rate of output - or maximum potential output in a period of time. When we speak about investments - money is used to generate more money. This is somewhat confusing. What is the capacity of $1M? What is the maximum output of $1M in a year? We don't know. But, we have a reference. Putting the money in the bank will make some money. Converting the 1M in a year to 365M dollar-days is correlated to the T generated by putting the money in the bank. All we need is to multiply this number by the daily interest. Well, not exactly. The impact of compounding interest are not reflected. Still, this is a reference.
When we're making an investment we lose the possibility to invest in other options (like the referenced bank account). Hence, both NPV and Flush are using negative sign for the investment as an expense. When a payment is received - we can invest again - which means we have again capacity to make money from money. When money is the constraint -- the constraining factor is how many constraining units you have at hand -- not any dollar-days counter. Suppose you have a great opportunity to invest 1.2 M for two years and then get 3M. If you have 1M at hand - you simply cannot make this investment. Time and money are NOT fully changeable. If your DD counter says you have 100 billion of DD - you still cannot make the investment - because you're constrained with dollars at hand. This is one of the missing factors both in the Flush method or the NPV. When money is a constraint - we need to know how much money is available for investment.
When two or more investments are possible - but, we're constrained by the cash (money - not money*time), we need to choose. Flush is trying to refrain from stating how much a dollar-day is worth. We give a value of money to everything we buy - but, we cannot give a value to being deprived of money for some time? Because of the reluctance to state the value of a dollar-day (a loan of a dollar for a day) - we need to assess the time all the invested DD will be returned, and this time is the deciding factor. The immediate consequence is that IT ISN'T RELEVANT WHAT HAPPENS AFTER THAT TIME PERIOD. You may get more payment then - or you got so much in the last day of the return that you can live forever on the interest alone. Flush will not relate to it at all. NPV will - in a very sensible way, assuming any dollar in the future is worth less than a dollar today - but, it worth something. This is not recognized by Flush at all. The time it takes to return the investment DD is an arbitrary number. I could have looked into the time it takes to return 110% of the invested DD (I want to make a profit out of my investment - isn't it?). Nothing really happens at that date. The capacity (the rate at which I can make money of money) is restored once I got my money back. The T that was generated is spread over a period of time. Sometimes the spread is over infinite period of time. Question is how I measure the impact of time on all the payments. NPV gives a sensible answer based on the price I give for having to depart from my money.
By the way - certainly Flush gives better judgment than return on investment (or the time it takes to return the money value of an investment). However, NPV gives better answer than both.
Tony, regarding your example of looking into 1M for 10 years, I need to know what happens in 10 years. If this is just a statement of a period - then after 1 year you'll still measure the capacity as the DD available for the next 10 years from that time.
We need to judge the rational of Flush as being the decision criterion for investments. This is the objective. In order to judge between two different investments - which have different quantities and timings of money invested and different back payments (time and money) we need to consider the price of temporary departure of our money. The way to make money out of money is by such a temporary departure. How much we want to get back to feel compensated. This is not a universal value. Different people may place different prices. NPV shows the real trade-off. Flush doesn't.
14 Mar 97 jonalisa@chesapeak.com
Have been following this whole thread with great interest. Thanks to all who are clearly investing quite a bit of their time writing their insights.I'm not going to comment on the whole thing, but I do have a question for Larry. Larry, you stated:
ow, let's follow Tony's idea, and say we have a pile of money. What is it? It certainly isn't T. It's not OE unless we use it for that. So it must be I.
OK, that makes sense.
When I put it into a Project, what happens? Doesn't it become OE?
Certainly if it is an internal product development. What if it is a contract project we are doing for someone else?
I don't understand this logic, Larry. If we agree that Inventory (I) is all the money that the system intends to turn into more money, and Operating Expense (OE) is all of the money the system uses to turn I into throughput (T), then it all depends on what you are spending this money on. I don't think that you can just say as soon as you claim the money is for a project, it becomes OE. It seems to me that it all depends on what you're spending the cash on, before you would claim it to be OE. In talking about a project, perhaps a different way to look at it is that cash is a raw material, and once it's invested in a project or an investment, the cash becomes work in process I. Money tied up in the system that is not yet turned into T.
14 Mar 1997 tocguy@lucent.com (Anthony R Rizzo)
Eli S. writes:I believe we can discuss the matter until we feel we understand the rational and the underlining assumptions.
Good! This is my goal for this discussion. I'm glad that we share this goal.
In my previous post I wrote: "When we speak about our time being a constraint --- this is NOT dealt by the Flush. The one constraint we deal with Flush is money".
Tony had reacted:
> I have a clarity reservation regarding this statement. Why do you say that time is not addressed by the flush calculation? I don't understand this.
I've tried to address a previous post by Tony, where he's described both money and time as OUR personal constraints as human being. How do you exploit your time is not addressed by Flush. How you treat your own money MAY be addressed by Flush.
Time in itself is not a constraint. It has to be linked to a constraint to be a capacity constraint. "We don't have time" implicitly says that a certain resource cannot produce what is required within the time frame.
The same can be said of anything that we consider a physical resource constraint. Just as time has to be linked to a resource, to be a capacity constraint, so too a resource has to be linked to time, to be a capacity constraint.
Consider a hypothetical race car that can travel at the great speed of 500 miles per hour. Does this car have the capacity to run the Indianapolis 500? It might, if it could sustain that speed for at least one hour. But if can maintain that speed for only a few seconds, then the car might not have enough capacity to finish even a single lap.
Flush (and NPV as well) is a method to measure the exploitation of money as a constraint. As Mark also states that Flush and NPV cannot be directly compared - let me submit that BOTH ARE USED FOR THE SAME PURPOSE - to set the investments priorities when money is constrained. When two different methods are supposed to do the same: COMPARING THE TWO IS INEVITABLE.
Clear.
In my previous post I've agreed to Tony's claim that the capacity of money is money*time. Thinking a little bit more has convinced me otherwise. Capacity is a rate of output - or maximum potential output in a period of time.
Well, which is it? These are two entirely different things. The former is a rate, a speed. The latter is the integral of that rate over a finite period of time. They are not the same thing.
I suggest that we define the term capacity as follows, at least for the duration of this discussion:
CAPACITY: The maximum potential output of a resources, over an available interval.
When we speak about investments - money is used to generate more money. This is somewhat confusing. What is the capacity of $1M?
Since my definition of capacity is really one of the two possible definitions that you suggested, I assume that you agree with my definition of capacity. The question, "What is the capacity of $1M?" has no meaning, in light of the above definition. The question can have meaning only in the context of time. The capacity of $1 Billion and a few nanoseconds is pretty small.Conversely, the capacity of $1 and a hundred years can be quite large.
What is the maximum output of $1M in a year? We don't know.
This question is different; it is entirely consistent with the definition; and it has meaning. Further, I suggest that the product, $1M times 1 year, is one measure of the "maximum potential output" of $1M over the year. Eli, are you trying to make my point?
But, we have a reference.
Exactly! We have a reference. In my opinion, the dollar-day calculation provides an objective reference that cannot be easily manipulated, to justify one's prejudices regarding a set of investments.
Does the NPV calculation provide such an objective, totally indifferent reference? Or do we tend to "adjust" the interest rate according to the "risk that we _feel_ exists?"
Putting the money in the bank will make some money. Converting the 1M in a year to 365M dollar-days is correlated to the T generated by putting the money in the bank. All we need is to multiply this number by the daily interest. Well, not exactly. The impact of compounding interest are not reflected. Still, this is a reference.
Clear.
When we're making an investment we lose the possibility to invest in other options (like the referenced bank account). Hence, both NPV and Flush are using negative sign for the investment as an expense.
Clear.
When a payment is received - we can invest again - which means we have again capacity to make money from money.
Ah! But what is the capacity of that new money? Again, the question is without meaning, in the absence of a reference to time. Without an available interval over which the investment can work, the very concept of capacity can't be used.
When money is the constraint -- the constraining factor is how many constraining units you have at hand -- not any dollar-days counter. Suppose you have a great opportunity to invest 1.2 M for two years and then get 3M. If you have 1M at hand - you simply cannot make this investment. Time and money are NOT fully changeable. If your DD counter says you have 100 billion of DD - you still cannot make the investment - because you're constrained with dollars at hand.
Let's put your statements in a logic cluster:
If I have on hand $1M,
and if I have available 100 billion dollar-days,
and if there exists an investment opportunity that requires $1.2M,
then I cannot make this investment.I have an entity existence reservation with the effect, "I cannot make this investment."
If I have on hand $1M,
and if I have available 100 billion dollar-days,
and if there exists an investment opportunity that requires $1.2M,
then I can very probably make the investment at a later date.Your assumption, of course, is that the investment won't wait.
This is one of the missing factors both in the Flush method or the NPV. When money is a constraint - we need to know how much money is available for investment.
I feel uncomfortable with this line of reasoning. We've been using the term, constraint, in the sense that a resource is too limited to do ALL that we might want it to do. But the underlying assumption has been all along that the resource is not too limited to do any one thing. We began this discussion from the perspective that we need to choose from a number of possible investments that we can make today. The operative word are "possible." We did not begin this discussion from the perspective of what's impossible.
Nevertheless, if there exist a sufficient number of dollar-days in my money*time critical constraint resource, then I can make even your $1.2M investment. Even this investment is possible. Albeit, it is possible only in the future. But it is possible for me to make this investment within the period of time that I have available for such investements.
When two or more investments are possible - but, we're constrained by the cash (money - not money*time), we need to choose. Flush is trying to refrain from stating how much a dollar-day is worth. We give a value of money to everything we buy - but, we cannot give a value to being deprived of money for some time?
I have an entity existence reservation with the statement, "Flush is trying to refrain from stating how much a dollar-day is worth." Quite the contrary is true. Flush says that the value of $1 for one day is far greater than the highly discounted value suggested by an assumed interest rate! This is quite different from saying that flush doesn't assign a value to a dollar-day.
Because of the reluctance to state the value of a dollar-day (a loan of a dollar for a day) - we need to assess the time when all the invested DD will be returned, and this time is the deciding factor.
Regarding the "reluctance [of flush] to state the value of a dollar-day...," please see the above entity existence reservation.
I have an entity existence reservation with the statement, "...and this time is the deciding factor." It is not! I've stated that the impact of the decision on the RESULTING CAPACITY of the investors money*time constraint should be the deciding factor. Under some circumstances, this RESULTING CAPACITY criterion and the time needed to reach the zero-flush point yield the same decision. But I've stated categorically that it would be very unwise to use the zero-flush interval alone, without questioning the underlying assumptions.
The immediate consequence is that IT ISN'T RELEVANT WHAT HAPPENS AFTER THAT TIME PERIOD.
I have a causality existence reservation with this conclusion. The claimed cause does not exist. Therefore, the conclusion is not valid.
You may get more payment then - or you got so much in the last day of the return that you can live forever on the interest alone. Flush will not relate to it at all.
I appears that by the statement "Flush will not relate to it at all," you mean that the time to reach the zero-flush point does not relate to it. If this is what you mean, then I refer you to the above discussion regarding the use of the zero-flush point alone, as a decision making tool.
NPV will - in a very sensible way, assuming any dollar in the future is worth less than a dollar today - but, it is worth something. This is not recognized by Flush at all. The time it takes to return the investment DD is an arbitrary number. I could have looked into the time it takes to return 110% of the invested DD (I want to make a profit out of my investment - isn't it?). Nothing really happens at that date. The capacity (the rate at which I can make money of money) is restored once I got my money back.
Here, you use the term capacity to mean a rate. I think that this is one source of confusion. I'll stick to the definition of capacity that I stated earlier. According to that definition, the available time interval must be taken into account. Indeed, once you get your original money back, the prior rate at which that money can generate more money is restored. But this rate is not a capacity. If, after you get your money back, you are able to invest for one additional year, then you have one capacity restored. If, after you get your money back, you are able to invest for one additonal day, then you do not have the same capacity restored. In fact, the word restored may not be the word to use here at all. A more appropriate phrase would be, "you have available one capacity or another." One of these might equal your original capacity.
The T that was generated is spread over a period of time. Sometimes the spread is over infinite period of time. Question is how I measure the impact of time on all the payments. NPV gives a sensible answer based on the price I give for having to depart from my money. By the way - certainly Flush gives better judgment than return on investment (or the time it takes to return the money value of an investment). However, NPV gives better answer than both.
I interpret your statement "the time it takes to return the money value of an investment," as the payback period. If my interpretation is correct, then I have an entity existence reservation with your last statement. Isn't the calculation of the payback period based on the same assumptions and on the same interest rate as the net present value calculation? Am I missing something here?
Tony, regarding your example of looking into 1M for 10 years, I need to know what happens in 10 years. If this is just a statement of a period - then after 1 year you'll still measure the capacity as the DD available for the next 10 years from that time.
No! This isn't just an arbitrary statement of a period. It is the ENTIRE period that the investor has available for investments. After 1 year, the investor has only 9 years left for such investments. Therefore, the capacity of the money*time resource is diminished by an amount that corresponds to 1 year.
We need to judge the rational of Flush as being the decision criterion for investments. This is the objective. In order to judge between two different investments - which have different quantities and timings of money invested and different back payments (time and money) we need to consider the price of temporary departure of our money. The way to make money out of money is by such a temporary departure.
Clear.
How much we want to get back to feel compensated. This is not a universal value. Different people may place different prices.
We agree here. There are many subjective measures available to us. Different people place different prices on the time-value of their money. This is reflected by the relatively wide range of interest rates that we encounter, when we try to get a loan.
NPV shows the real trade-off. Flush doesn't.
We don't agree here. But maybe that's because you might be talking about the time required to reach the zero-flush point, while I'm talking about the REMAINING combined capacity of an investor's money and time.
When all is said and done, I think that I've understood your thinking a little more clearly, and we're still friends. What more could I ask for? :-)
14 Mar 1997 tocguy@lucent.com (Anthony R Rizzo)
Eli writes:>Tony, regarding your example of looking into 1M for 10 years, I need to know what happens in 10 years. If this is just a statement of a period - then after 1 year you'll still measure the capacity as the DD available for the next 10 years from that time.
I responded:
No! This isn't just an arbitrary statement of a period. It is the ENTIRE period that the investor has available for investments. After 1 year, the investor has only 9 years left for such investments. Therefore, the capacity of the money*time resource is diminished by an amount that corresponds to 1 year.
However, the alternate response would be correct too. Even if this were "just a statement of a period..." after 1 year one would not still measure the capacity as the dollar-days available for the next ten years from that time, because one would want to use a constant future date as the fixed reference against which one would compare all the candidate investments.
The problem with the flush calculation, all along, has been the open-endedness of the calculation. Consider the case that you suggested in an earlier article. I invest $1M today, and I get $2M in a year. By the current definition of flush, I don't have to do a thing, in order for that $2M to generate "flush" for me. This is where we typically have difficulty with the concept. The accumulation of "flush" without making further investment doesn't make any sense at all.
But this difficulty is caused by a wrong sign. That wrong sign, in turn, is caused by our use of the investment date as the reference date and by our lack of use of a fixed-size reference interval. All this clear up, when we think in terms of a money*time resource, where the money component is the amount of money on hand at the time of investment, and the time component is an interval of fixed duration.
Consider, again, the same example that we discussed earlier. The investor has available $1M and 10 years. The investment is $1M now. The payback is $2M in 1 year.
By the current flush calculation, flush would be:
At time zero: Flush = zero
At time 1 yr: Flush = -($1M * 1Y),
This is the time when we get the payback of $2M, but the current flush calculation doesn't do anything with this payback at this time.
At time 2 yr: Flush = -($1M * 2Y) + ($2M * 1Y) = zero
At time 3 yr: Flush = -($1M * 3Y) + ($2M * 2Y) = 1 M$Y
At time 4 yr: Flush = -($1M * 4Y) + ($2M * 3Y) = 2 M$Y
Flush continues to increase with time now, but the investor has done nothing with the $2M earned from the investment. Clearly, this doesn't make any sense.
Now think in terms of calculating the capacity of the investor's money*time resource. So as to not confuse the current definition of flush with the concept of a money*time resource, I'll use the term MoneyTime to refer to the resource. Recall that we stated that the investor had available 10 years.
At time zero: MoneyTime = $1M * 10Y = 10 M$Y
This is the capacity of the resource at time zero.
At time 1 yr: MoneyTime = ($1M * 10Y) - ($1M * 1Y) + ($2M * 9Y) = 10 - 1 + 18 = 27 M$Y
This is the entire MoneyTime resource available to the investor after the first year. Notice that we now have to take the earnings into account as soon as they occur. Conceptually, this makes great sense. After all, the investor now has those earnings on hand. They clearly increase the capacity of the MoneyTime resource.
At time 2 yr: MoneyTime = 27 M$Y - ($2M * 1Y) = 25 M$Y
In other words, the investor simply held onto the $2M for a year, and did nothing with it. Therefore, the capacity of the MoneyTime resource is diminished by an amount equal to ($2M * 1 year).
At time 3 yr: MoneyTime = 27 M$Y - ($2M * 2Y) = 23 M$Y
At time 4 yr: MoneyTime = 27 M$Y - ($2M * 3Y) = 21 M$Y
In other words, if the investor does nothing with the $2M earned at the end of the first year, then the MoneyTime resource is depleted by an amount equal to the product of cash on hand and time wasted. If the investor does no more investing, then at the end of the 10 year period he/she will have on hand $2M. But the capacity of the resource will be zero, because the time that the invetor had available for making investments is gone.
Doesn't this make more sense?
15 Mar 97 Eli S <elyakim@netvision.net.il>
Tony has responded in two different posts to my last on regarding: Flush or NPV.I have defined capacity in this way:
>Capacity is a rate of output - or maximum potential output in a period of time.
Tony responded:
Well, which is it? These are two entirely different things. The former is a rate, a speed. The latter is the integral of that rate over a finite period of time. They are not the same thing.
Capacity is a rate: d(output)/dt. Assuming this rate is approximately constant over a certain period of time we may approximate it by: output/time. We do it for speed: how many miles did you do in an hour.
Capacity is defined as a rate. The formal definition by APICS is: The highest reasonable output rate which can be achieved with the current product specifications, product mix, work force, plant, and equipment.In the second response by Tony, he refers to it
All this clear up, when we think in terms of a money*time resource, where the money component is the amount of money on hand at the time of investment, and the time component is an interval of fixed duration.
I agree to that. The interval should be fixed for the capacity calculation. That means capacity is a rate, and we can approximate it by refering to the money*fixed-interval-of-time.
Tony also says:
In my opinion, the dollar-day calculation provides an objective reference that cannot be easily manipulated, to justify one's prejudices regarding a set of investments.
Does the NPV calculation provide such an objective, totally indifferent reference? Or do we tend to "adjust" the interest rate according to the "risk that we _feel_ exists?"
Tony has a point. Dollar-days is objective while the inclusion of the interest rate is subjective. But, it seems to me that preferences of investments have to include that subjective judgement. The risk and the benefit of having $1M NOW versus having $1.5M in the future is subjective. If I don't convert the DD into money I may get a decision that will contradict my intuition. Ang my intuition may be right! The negative branch of the tendency to justify one's prejudice is certainly a negative branch. It needs to be addressed. I don't claim that NPV is perfect! One needs to go into it and make it better. As a matter of fact, Tony suggests to introduce changes to Flush. For instance, he suggests not to rely on the time it takes to return the DD of an investment as the criterion for decision -- but, on capacity estimation for some future date.
The rational of Net-Present-Value (NPV) can be applied to bring every money transaction to a future date. It will generate the same decision criterion as bringing it to the present date.
The following is, to my mind, especially illuminating:
Let's put your statements in a logic cluster:
If I have on hand $1M,
and if I have available 100 billion dollar-days,
and if there exists an investment opportunity that requires $1.2M,
then I cannot make this investment.I have an entity existence reservation with the effect, "I cannot make this investment."
The reservation is because in the future I'll have $1.2M and will be able to make the investment. That's true - depending on the assumptions about the availability of that investment. Still, the constraining factor is the amount of cash - not the DD in the counter. As Thoughput is, in this case, added capacity -- what I cannot do today I may be able to do tomorrow. However, the future investment needs to be done in the future - so, its impact should be reduced.
This can lead to an observation about dollar-days versus NPV. Suppose in the life of the project you'll need to give away $1M and get it back after 10 days. Suppose you can do it now --- or you can do it one year from now.
From DD point iof view there is no difference. From NPV there is a difference: better to do it next year --- because the value of money next year is less than it is now. I think this is reality. What I can have now worth more than the same amount next year.
I have an entity existence reservation with the statement, "Flush is trying to refrain from stating how much a dollar-day is worth." Quite the contrary is true. Flush says that the value of $1 for one day is far greater than the highly discounted value suggested by an assumed interest rate! This is quite different from saying that flush doesn't assign a value to a dollar-day.
I don't understand this argument. In what sense Flush is claiming that 1 dollar-day is greater than the interest-rate (which may be very large). Money is used to compare different benefits/values. Flush doesn't convert dollar-day to money --- so, it can compare it only to dollar-days. When different options regarding time exist -- I need to weight the difference: how much worth $1 today than $1 next year. As this difference means the benefit of having something NOW versus having it in the future -- it is an entity that has a price! By the way, from Flush point of view - both go to infinity as no return is considered.
In his later post Tony suggests a new way to make a decision based on DD:
The problem with the flush calculation, all along, has been the open-endedness of the calculation. Consider the case that you suggested in an earlier article. I invest $1M today, and I get $2M in a year. By the current definition of flush, I don't have to do a thing, in order for that $2M to generate "flush" for me. This is where we typically have difficulty with the concept. The accumulation of "flush" without making further investment doesn't make any sense at all.
But this difficulty is caused by a wrong sign. That wrong sign, in turn, is caused by our use of the investment date as the reference date and by our lack of use of a fixed-size reference interval. All this clear up, when we think in terms of a money*time resource, where the money component is the amount of money on hand at the time of investment, and the time component is an interval of fixed duration.
This makes more sense - but, it is practically what NPV is doing. As I've already said - you make the calculation of Future Net Value and get the same priorities between investments. Conceptually there should be no difference if you bring all money transactions to the present or to the future. The difference between DD (dollar-days) and NPV lies in two factors:
1. The DD doesn't give a money-value to a dollar-day. In this sense it is more objective.
2. The DD calculation doesn't lend itself to compound interest.In reality - when you put your money in the bank you get interest on your interest's earnings. But, in order to consider it - you need to decide what are the interest-earnings.
The basic assumption behing DD and Flush is that if you have money you have some benefit (security - for instance) OR you can invest it in a simple and structured way. But, the impact of compounding interest is ignored in the DD calculation and it is typical to the reference for investment: depositing the money in a bank and gaining interest over time.
This long discussion demonstrates that the Flush solution is not simple to understand - even though it is easy enough to calculate. For me NPV is much easier to understand. It is more difficult to calculate. But, computers can do the calculation. The concept of the value (in money) of lending $1 for 1 day is, for me, simple. The rest is build upon this.
The TOC education takes care to show the flaws of any current algorithm that TOC suggests another way. The traditional cost, for instance, occupies a great deal of writing and teaching to show where it goes wrong. When Flush was introduced - what attack on the NPV or IRR was done? One has to acknowledge that there are old ways to do what Flush is supposed to do - - and see why is the new way better than the old. Flush was introduced in 1989.
I'm not aware of any material to cover the topic of this discussion.
15 Mar 1997 lleach@srv.net (Larry Leach)
Mark,>My two cents is that if we focus on the time=money issue, we miss the point.
Flush, as described by its proponents, does not in any way suggest time = money. It is a variable in units of money*time. This is entirely different.
Back to the example, torque is measured in units of force*distance. There is no assumption that force = distance. It is true, however, to balance torque, the force*distance products must be equal.
If this doesn't explain it, I need to think up another example.
Perhaps this is one of the perceptual blocks.
>I think the only constraining resource is money.
Yes, Flush was only proposed for cases when money is a constraint.
I'm thinking that time is ALWAYS the ultimate constraint. It cannot realistically be lengthened (as far as I know). Therefore, it need not be considered a constraint separately.
Clarity?
My point was that we are trying to choose the "best" solution among multiple options - regardless of how we come to that conclusion (what is "best"). Thus, an option that returns money sooner or at a higher rate over the long term is the "best" solution (check your goal assumption here:))
For sure.
The "best" scenario is one that gets us more throughput in the least amount of time with the least amount of investment.
Sure. The problem we are working is, "Does NPV give us the information we need to properly make that judgment, or is another tool needed?"
16 Mar 1997 tocguy@lucent.com (Anthony R Rizzo)
Eli, [d]uring the course of our discussion, you've stated that you can think of countless cases where not having sufficient cash on hand can block an investor from making an investment today. I certainly can see your point. Without enough money on hand now, an investor can't invest in many instances. That's clear.If I had an entity existence reservation regarding your claim that money is often a constraint, then a simple predicted effect test would prove my reservation invalid:
If money is almost never a constraint,
then an investor can make nearly any investment at any time.Clearly, the predicted effect does not exist. Therefore, the cause cannot exist. Further, since the stated cause is mutually exclusive with the claim that money is nearly always a constraint, then the statement that money is nearly always a constraint must be accepted as true. So, certainly, money is most often a constraint.
However, I, in turn, can think of coutless cases where a shortage of time can block an investor from making progress toward the investor's goal. Can we make use of an investment that pays off in a year, if we have but a week available? Let's see if the same predicted effect test works in this case:
If time is usually not a constraint,
then an investor often can wait as long as necessary, to make most investments.If an investor often can wait as long as necessary, to make most investments,
and if an investor can make many intermediate, smaller investments successfully,
then an investor can make nearly any investment with very little money.Again, the predicted effect does not exist. Therefore, the cause cannot exist. Further, since the stated cause is mutually exclusive with the claim that time is often a constraint, then the statement that time is often a constraint must also be accepted as true.
From where I sit, it appears that we're both right. Money can be and often is a constraint. Time, too, can be and often is a constraint.
When we began this discussion, I wasn't completely certain that the product of money and time could be considered a critical resource. Now, I have no doubt. The product of money and time is a measure of an investor's critical constraint resource. But the interval that represents the investor's available time, of course, depends on the investor, because every investor has in mind some latest possible date beyond which the very act of investing makes no sense. Do you know of any investor who can afford to wait forever, for an investment to pay off? Neither do I.
The realization that the product of money and time can be considered a critical resource, combined with the paradigm shift that I made when I learned of the Theory of Constraints and the five-step process of ongoing improvement, cause me to want to exploit my own critical constraint resource, which I call MoneyTime. They also cause me to want to subordinate all my other investment-related decisions and processes to my decision to exploit my MoneyTime critical constraint resource. I think that I shall do so, in the future.
When we began this discussion, I had a few concerns regarding the flush calculation. For example, the accumulation of flush in the absence of additional investments didn't make sense to me at all. I doubt that it made sense to any one. But this discussion has provided significant clarity for me. Even the flush calculation clears up, at least in my mind, when I think in terms of a MoneyTime critical constraint resource.
The point of this discussion wasn't for me to convince you or for you convince me. It was to achieve a clearer understanding of each other's thinking. It was also to achieve a clearer understanding of our (my) own thinking. From this side of the world, it appears that the objective of clarity has been achieved. For this, and for the kindly and professional manner in which you've put forth your ideas, I thank you.
97-03-17 Mark_Woeppel@msn.com (Mark Woeppel)
Larry,Thanks for the response. I wrote:
>My two cents is that if we focus on the time=money issue, we miss the point.
To which you responded:
Flush, as described by its proponents, does not in any way suggest time = money. It is a variable in units of money*time. This is entirely different.
Yep.
Back to the example, torque is measured in units of force*distance. There is no assumption that force = distance. It is true, however, to balance torque, the force*distance products must be equal.
If this doesn't explain it, I need to think up another example.
Perhaps this is one of the perceptual blocks.
No, I understand completely.
I think the only constraining resource is money.
Yes, Flush was only proposed for cases when money is a constraint.
>I'm thinking that time is ALWAYS the ultimate constraint. It cannot realistically be lengthened (as far as I know). Therefore, it need not be considered a constraint separately.
Clarity?
Suppose time is not a factor. I have an infinite amount available. Does this change our evaluation? It may or may not. However, our judgement of the relative usefulness of a particular measurement (as we are discussing here) without a time component will be that it does not meet our needs. In other words, there must be another component of the measure that gives it meaning.
If we look at the global measurements used to judge effectiveness of our enterprise, all have a time component. Throughput = rate at which revenue is generated (per unit of time). Operating expense, all the money spent to turn inventory into throughput (per unit of time). Inventory is an absolute number, but we use time to give it meaning; i.e., ROI (per unit of time). Turns is the use of inventory as a function of time.
The "best" scenario is one that gets us more throughput in the least amount of time with the least amount of investment.
Sure. The problem we are working is, "Does NPV give us the information we need to properly make that judgment, or is another tool needed?"
If this is the debate, then I have just chased a rabbit down its hole, and I'm way off track. I was responding to a particular statement about what is the constraint using Flush. Is it time or money or a combination?
I was weighing on the part that time is not the constraint, but money is. The time component is "oxygen" that need not be stated explicitly as the constraint. So - it seems like we drifted off into an arcane discussion of angels on pins.
97-03-18 jeaster@verinet.com (James E Easter)
From: Anthony R Rizzo <tocguy@lucent.com> who wrote:When we began this discussion, I had a few concerns regarding the flush calculation. For example, the accumulation of flush in the absence of additional investments didn't make sense to me at all. I doubt that it made sense to any one. But this discussion has provided significant clarity for me. Even the flush calculation clears up, at least in my mind, when I think in terms of a MoneyTime critical constraint resource
I am pretty pleased with the amount and depth of discussion on Flush and NPV which has flowed since my question several weeks ago. Thanks to all who have spent time thinking and writing on this topic. Since I started this whole thread about NPV and Flush, let me ask again my original question (paraphrased) and get clarity on the latest answer:
* Under what scenario would using NPV (vs. Flush) lead me to make an erroneous investment or decision? (I make an investment I shouldn't have) when using the FLUSH measurement would have lead me to avoid or not make the investment?
and
* conversely, when would the opposite be the case--that is, using NPV would cause me NOT to invest and FLUSH would have me do so.....
To avoid simply soliciting the exact same responses to date, let me add some assumptions:
- I have enough cash to make the investment(s)
- I have enough time to wait for the investment to run its course
The general scenario is the one we are often faced with:
We have more than one option for our cash or line of credit, etc. It may be two competing projects, or simply an idea for a project vs. do nothing. In all cases we have two or more choices for our cash, LOC, etc.
I asked these questions precisely--because the ultimate characteristic of a good "measurement" is how reliable it is in inducing me to make good decisions (and behaviors)--- I want to uncover the (potential) gem that Flush may offer over NPV....if there is one.
97-03-19 jonalisa@chesapeak.com
Hi Larry,I'm going to respond by answering your last question first, OK? Your question was, "I am not sure I understand the idea of cash as raw material. Can you clarify?"
Gladly! First, let's examine a business system that accomplishes its purpose by manufacturing stuff. The macro process looks something like this: Materials get purchased from vendors and are introduced into the system as "raw materials inventory." The manufacturing system takes those raw materials through a conversion process. From the moment the manufacturing system begins that conversion process, until the point at which that process is "complete," those raw materials are reclassified as "work in process." Once the manufacturing system completes its conversion process, the result is a "sellable" product, and the inventory is reclassified again as finished goods.
But, especially when we're talking about a business system that has as its purpose, "to make more money now as well as in the future," we must shift our thinking a bit. That stuff, whether we call it raw materials, work in process, or finished goods, is really MONEY. All these business systems are really doing, hopefully, is continually converting piles of money into bigger and bigger piles of money in measured periods of time.
Our conversion process began not when we started to manipulate the raw material "stuff," but when we converted that cash into a specific something that we purchased from a vendor. My claim is that cash is the base raw material of every such business system -- raw materials originate in your bank account, not in your stockroom.
The main difference between money that gets classified as inventory and money that gets classified as operating expense, is that the money that's classified as inventory is money that is still captured within the system (regardless of whether it looks like cash or whether it looks "stuff."). Money that gets classified as operating expense is money that is spent for the purpose of operating the system, but the cash, once spent, is no longer money "in" the system.
Let's take a look at a scenario different from a manufacturing system. You purchase a new car. The price of the car is $25,000. You are going to pay cash for the car. So, you go to the bank and pull out $25,000 from your account, and then hand that $25,000 over to the car salesman, who hands you the keys to your new car. At that moment, you have transformed your $25,000 cash-inventory into $25,00 car-inventory. The next day, the car depreciates by $3,000. Your car-inventory is now $22,000, as it is still your money tied up in your system. Basically, you have "spent" $3,000 of your $25,000 car-money. This $3,000 is operating expense -- it's no longer "your" money.
I hope I've been clear so far, Larry. Please let me know.
OK, Larry, now lets look at the two projects you are contemplating. I guess I'm going to continue with my last-in-first-out approach to your message here, and address the second type of project first.
"The other kind of project I was thinking about was when a company does a project on a contract basis; e.g., builds a road or helicopter or something. In these cases, the product is not an asset to the company. Yes, there is raw material cost. And I guess we could make it work in process I, but that is valued at the raw material cost in TOC accounting, isn't it?"
Yes. The reason I agree with you here is that I can view this "project" in the same way that I view manufacturing. I think the main reason we call it "project" as opposed to "product manufactured by the company" is that the lead time tends to be so long on something like a road or helicopter. The result of the project is what the customer is buying from the company.
"I was thinking most of the work performed by the company is OE"
This seems to make sense to me, too, assuming that by "work" you mean the salaries, etc., of the folks who are doing the work.
"The first are internal projects like product development. I think most of that 'investment' really becomes OE as we perform the project; maybe a little bit of I is created in terms of prototype manufacturing equipment."
As is the case with the TOC measures in manufacturing, perhaps we should differentiate between the measures that must be used for reporting purposes -- like to "corporate" entities and the government -- and measures that should be used for reality-based decision making. The latter is what I would like to discuss for two reasons. First, I don't believe it's as straightforward as you suggest. Second, I don't have the answer! This issue of "R&D" has been the subject of much discussion over the years. Our quest is to determine how to us T, I, and OE for projects like this so that the measures reflect reality (a). The question usually boils down to this conflict: "treat the salaries/fees of the folks dedicated to the R&D effort as OE (d)" versus "treat the salaries/fees of the folks dedicated to the R&D effort as I (e)". We aren't sure whether to treat these reclassifications of our cash as investment or expense. Certainly, we are paying these salaries/fees on a regular basis. And, this money is converted into "ideas" or "knowledge" -- how the heck do you attach any kind of realistic "value" to ideas and knowledge? This leads us to the inclination to treat them as OE. But, at the same time, these salaries/fees are devoted to long term issues. We don't see the efforts turn into throughput for some time. And, the ideas and knowledge should be worth something, shouldn't they? This leads us to the desire to treat the salaries/fees as inventory. There are potential ramifications of either approach, when we remember that measures are a driver of behavior.
Hey, you're talking about three types of projects, you sly fox, you!
"I wasn't thinking about pure capital upgrade projects. These will become I. I am not sure when, though. For example, in the Utility industry law requires that they can not call a new power plant a capital asset till it is on line. Therefore their payments are OE till it is on line. I am too out of date on accounting to know about other industries."
I am assuming that they should do what the law requires -- this is a necessary condition. But, how SHOULD they account for this stuff for business management and decision making purposes? To be honest and selfish, my interest is in the above R&D issue more than this one. So, I'll leave this discussion to others.
97-03-21 caspari@iserv.net (John A. Caspari) Subj: FLUSH (interest and NPV explained)
On 13 Mar 1997 lleach@srv.net (Larry Leach) wrote:... What is the 'real' cost of this money?
Don't see any 'interest' involved in the calculation at all, so it id difficult to see how interest or interest rates should have anything to do with it.
And on 9 Mar 1997 tocguy@lucent.com (Anthony R Rizzo) wrote:
If (300) The zero-flush interval is the interval required for an investment to completely releases an investor's critical constraint resources,
and if (310) for some investments the zero-flush interval is infinite,
then (320) some investments never completely release the investor's critical constraint resources.These two postings have caused me to think that additional clarity may be useful for the FLUSH discussion. When I first encountered FLUSH it was presented as a measure which would aid us in avoiding making investments such as Tony describes in entity 320. It was contrasted with Net Present Value ("NPV"), which was illustrated as showing the investment to completely recover its investment cost. I was uncomfortable with the presentation because it appeared to me that what was described as "NPV" was, in fact, simply an undiscounted payback period. Today I have the uncomfortable feeling that a part of the FLUSH versus NPV controversy may due to a semantic difficulty relating to the definition of 'NPV'. I will describe NPV briefly to be sure that we are all talking about the same thing.
Addressing Larry's comment regarding the cost of money and interest: the 'real' cost of money, as portrayed in the financial management literature, is the cost of capital for the organization (the cost of acquiring additional money or the cost avoided by needing to acquire less money). This is an opportunity cost measure and represents the alternative to using the money in the business. The decision principle is that if an organization can earn more than its cost of capital on an investment, then it should make the investment. If the investment returns less than the cost of capital, then the investment should not be made.
As instructors, we often try to illustrate this concept with a simplified example. Assume that you are currently paying 18 percent interest on a credit card outstanding balance of $3,000. You receive an inheritance of $1,000. You have the opportunity to make an investment which returns 15 percent. Should you make the investment or pay off some of the credit card debt? Would your answer be different if the investment returned 25 percent? In this case, we are assuming that the 18 percent credit card rate is the cost of capital for the individual. Of course, as Eli Schragenheim has observed, the cost of capital concept (which includes interest as a part of the calculation) is somewhat more complex than this. Because of this sort of example, and also the fact that the discounted cash flow (DCF) analysis methods are based on the "compound interest" formula [(1+ (interest rate))^(number of periods of compounding)], we often use the term, 'interest rate', to refer to the DCF discount rate. But understand, we are talking about an appropriate discount rate, not an interest rate.
Now, as to NPV. We discount a stream of cash flows projected for an investment proposal at a discount rate which is equal to the organization's cost of capital. The result will be positive if the proposal returns an amount greater than the discount rate. The result will be negative if the proposal returns an amount less than the discount rate. Thus, the sign of the NPV may be used to determine if an investment proposal is acceptable at all.
NPV is not used to compare two acceptable and mutually exclusive alternatives. The NPV technique for this is the Profitability Index. The Profitability Index states the present value (discounted at the cost of capital) of the future net cash inflows in terms of the investment amount required [(PV of net cash flows) / (investment required)]. A profitability index greater than 1.0 implies a positive NPV. The investment proposal with the larger positive profitability index is preferred. In using the profitability index to decide among competing proposals, it is implicitly assumed that reinvestment and returns during unequal economic lives are at the cost of capital. In some cases there is a problem in computing the amount of investment versus the future cash inflows for use in computing the profitability index.
A second discounted cash flow method is known as the Internal Rate of Return (IRR) or Time-adjusted Rate of Return. The IRR is the rate of return of the proposal. A proposal is acceptable if the IRR is greater than the cost of capital. In comparing mutually exclusive proposals, the acceptable proposal with the larger IRR is preferred. This method implicitly assumes that reinvestment and returns during unequal economic lives are at the rate of return of the proposal with the higher IRR. It is not necessary to separate the investment amount from future cash inflows when computing the IRR.
There is a relationship between the NPV and IRR of a proposal. The discount rate which yields a NPV = 0 is the IRR.
Are these (NPV, profitability index, and IRR) the discounted cash flow concepts that we are all using in the FLUSH versus NPV discussion?
97-03-21 caspari@iserv.net (John A. Caspari) Subj: FLUSH (Money as "I")
On 13 Mar 1997 lleach@srv.net (Larry Leach) wrote:1. We have learned, and presumably all agree on the TOC principles of T, I, and OE.
I have not seen these "principles." Are they enumerated somewhere? Or are we talking about the fact that we want T to increase, while decreasing I and OE?
[Larry responded that he was talking about the paradigm shift in the ranking of importance of T, I, and OE -- i.e., that T, not OE, must be viewed as most important.]
Now, let's follow Tony's idea, and say we have a pile of money. What is it? It certainly isn't T. It's not OE unless we use it for that. So it must be I.
I don't follow this reasoning. If money is I, then whenever we have more T (money generated) we also would have more I in exactly the same amount. Similarly, whenever OE is reduced (less money spent) we also would have an increase in I in exactly the same amount. An example: T is up by $1,000,000, I is up by $1,000,000, and there is no change in OE. An evaluation of the three measurements (T, I, & OE) would not give a clear picture at all.
I suspect that there is some amount of money that is "captured" by the system -- as a part of working capital -- that could be classified as I. The remainder of the money is just "M", neither T nor I nor OE. The "M" is available for distribution to shareholders (as either dividends or buy-back of stock) or for additional investment as new I.
97-03-21 jeaster@verinet.com (James E Easter)
Are these (NPV, profitability index, and IRR) the discounted cash flow concepts that we are all using in the FLUSH versus NPV discussion?I think I started this whole thing about NPV vs. flush--and John's definitions are a very good summary of what I meant by NPV--I was using the term to speak of using discounted cash flow methodology-including NPV, IRR, etc.--Nice job- John -of summarizing and defining terms.
So, yes, that is what I meant by "NPV". And, what is the answer?
97-05-02 tony@ermapper.com.au (Tony Clark)
I've just finished reading my copy of Critical Chain (things take a while to get to here) and have been puzzled by the concept of Flush used as a method of evaluating projects.Somehow, there seems to be something wrong with it.
Suppose I had 100 dollars with this opportunity to invest it with the largest bank, guaranteed by the FDIC, absolutely no risk. What's more, as a promotional offer they are prepared to pay me 25% interest for 1 year. Guaranteed, no catch, no risk. To me, it seems like a great opportunity. But let's evaluate it using flush. My investment is $100 for 365 days, or 36,500 dollar-days. Back end I get 150 dollar days. Doesn't sound so great, does it. Maybe, to come out even, I have to get back 36,500 dollar days. The only places I've heard of that sort of return are the fields of Colombia or the Golden Triangle.
In fact the evaluation based on $36,500 payback is exactly equivalent to a NPV with a discount rate of 36,400 percent.
So, maybe there's something wrong with my thinking here. Oh yes, maybe I have cut off the benefit too early. I get my $100 back, which recovers the original investment; but I also have $25 in interest which I enjoy for the rest of my life. No - it's even longer than that. I pass it on to my children or to the Government in taxes and death duties. So the benefit of that $25 goes on to infinity, in dollar days.
But there's something wrong with that. On that basis, I could get 1c back in interest, which would still be a great deal if I took dollar days to infinity. So why doesn't it feel so great?
Maybe we have to cut back the length of time we enjoy the benefits.
But let's get back to the question of Flush v NPV. As I mentioned above, a Flush evaluation showing breakeven after 1 year on an Investment of $1 would require a return of $365, yielding net dollar-days of zero. Equivalent to a NPV return of $0 with a discount rate of 36,400 percent
NPV = -1 + 365 / (1 + 36400/100) = 0
I suspect that taking longer periods will not change things.
Then I thought, "Why are we evaluating projects anyway? Are there any basic assumptions I'm making?" It turns out that there are.
- There is more than 1 investment opportunity in the period between money out and full recovery of that money.
- We have enough money to invest in at least one opportunity.
- We do not have enough money to invest in all the opportunities we identified in 1. above.
I also have convinced myself:
- That the methodology of Flush is equivalent to NPV at 36,400 percent discount rate
- That Flush, at the limit of enjoying the benefits of the investment, goes to infinity. Infinity is not a nice concept to use in comparing projects. On the other hand NPV handles the infinity question quite well. Discounting constant (or constantly increasing) cash flows to infinity leads to a finite value.
- That Flush will result in a huge bias towards early payback - at the expense of potentially vastly more profitable projects which derive cash flows later.
So I still lean towards NPV. But NPV to me has some major drawbacks. I personally don't like the way it handles risk (by upwardly adjusting the discount rate). My preference (which is probably not supported from a purely technical viewpoint) is to first establish a discount rate. Our goal is to make money for our shareholders. Our shareholders expect to receive a certain return on their money. If we don't deliver that, then we should be giving them the money back, not investing in projects. For me, then, this is the base discount rate. (Not the cost of money, as is argued in the literature. There is a condition that the cost of money must be lower than this rate). I then like to discount the investment separately from the resulting cash flows. By comparing the resulting net present values, I can then evaluate the margin available to cover risk. Riskiness is always a subjective assessment and I don't like the way it can get buried in the interest rate in a standard NPV calculation.
To me, the Flush method is too close to the payback method.
Go to TOC-L/Goldratt Postings Contents at http://users.aol.com/caspari/toclist.htm