Editor’s note: This article is the second of a two-part series entitled “Marginal Thinking.”

Calculating marginal feed costs
Few dairies actually know their marginal feed costs; amazingly, many think they know their average feed cost per hundredweight (cwt) of milk, and almost all know their daily feed cost per group. Although it is not exact, using average daily feed costs for pen groups allows an estimate of the marginal feed costs. Our goal is to estimate two values:

•the average maintenance cost per cow per day
•the additional feed cost per cwt of marginal milk

Note the first is per cow per day and the second is per cwt, regardless of how many cows or how many days it takes to produce that extra milk.

Let’s look at a hypothetical example of the relationship of production and average cost of production. Given the cost of feeding the cow producing 40 kilograms or about 88 pounds of milk ($4.50) and the cost of feeding a dry cow ($1.50), the cost of feed to support the 88 pounds of milk production can be calculated ($4.50 - $1.50 = $3.00). If it costs $3 to support the milk production of a cow making 88 pounds of milk, then each kilo must cost 7.5 cents for the feed.


Using the maintenance cost of feed for the dry cow as an estimate for all maintenance costs ($1.50) and knowing the feed cost of a marginal kilogram or 2.2 pounds of milk (7.5 cents), the cost of feeding at each level of production can be calculated. Given a price for milk, revenue and income-over-feed costs can also be calculated.

The line for average feed cost per 100 kilograms or 220 pounds is most telling. Notice that in a herd with a single TMR fed to all cows, the “average” feed cost per cwt varies hugely across production levels. Average feed cost per kilogram is a terrible measure of the economic efficiency of a dairy’s feeding management (e.g., purchasing cost control, inventory management, ration balance, feed delivery, etc.). Average feed cost per 220 pounds is almost entirely a proxy figure for the herd’s production level.

As mentioned earlier, it is easier to determine the herd’s level of production by simply dividing milk sold by cows milking. Average feed cost per 220 pounds provides almost no information regarding how well management is making feeding decisions.

Unfortunately, a similar measure gained popularity in the press – feed efficiency. It is not possible to determine the origin of this mistake, but it probably relates to the simplicity of calculation. Feed efficiency is merely the amount of milk divided by the amount of feed. However, this suffers from the exact same issues as average feed cost per 220 pounds of milk. As a matter of fact, it is the same calculation, merely adjusted by feed cost per kilograms of feed.

So, as we now understand, daily feed is really two components – maintenance feed and marginal feed. A high-producing cow has very similar maintenance feed to a low-producing cow. But that maintenance feed is a far greater proportion of the diet of the low cow. So, big surprise, a low cow has lower feed efficiency.

In simple terms, she is not diluting her maintenance as well as the high-producing cow. This feed efficiency is merely a function of milk production. As enticing as it might appear to certain people, the use should be banned on knowledgeable dairy farms.

Never look at feed efficiency or average feed costs as a measure of the efficiency of a feeding program; look only at marginal returns to make decisions. Average cost per ton of soybean meal may be a useful number in determining if the farm is doing an efficient job in purchasing feed. Average feed cost per hundredweight, in contrast, is so confounded by the herd’s production that it is misleading, at best, as a measure of the farm’s efficiency in feed procurement and delivery.

Feed cost per hundredweight on most dairies is something of a foregone conclusion. If the herd is producing poorly, the average feed cost per hundredweight will inevitably be high. If the herd is filled with high producers, the average feed cost per hundredweight will inevitably be low. Average feed cost per hundredweight is not the best monitor of how well a farm controls its feed purchase costs.

Calculations based on the benchmarks of “good cost control” versus “poor cost control” farms can show that more than one-half of the apparent losses to herds attributed to poor cost control in purchasing feeds are not, in fact, due to the price of the feed purchased. Instead, these losses are due to low production and failure to dilute out maintenance costs.

Clearly, the most effective way to reduce average feed costs per hundredweight is to feed and manage cows so they produce more milk. The marginal cost of feed is so small (about 4 cents per 1 pound of milk) compared to the marginal value of milk (12 to 15 cents per 1 pound) that it is essentially always a good idea to feed more to make more milk. When the milk price is low, the proper behavior is to try to make more milk per cow. When the milk price is high, the proper behavior is still to try to make more milk per cow. It is pretty simple.

While it is certainly true that dairy farmers should try to purchase their feeds as cost effectively as possible (buying feeds cheaply), it is probably rarely wise to buy poor feeds because their price is low (buying cheap feeds). The impact of marginal milk on profit is simply too great to forego. Veterinarians and nutritionists need to help their clients refocus away from feed cost savings and toward increased production and profit.

Does a low-production group save money by reducing feed costs?
Group-feeding cows by production group is so widely practiced in the dairy industry that it seems like common sense that it must be a wise management practice. If it costs an average of $4 per day to feed a cow in the high group and $3 per day in the low group, it would seem only to make sense to move cows to the cheaper ration as their production declines in later lactation and save the $1 on feed costs. If the average cow in a 100-cow dairy spends on average only the last 100 days in the low group, the savings should be $10,000 per year in feed costs (100 cows, 100 days each, $4 - $3 = $1 per day).

As appealing as this arithmetic may be, it is not true. Several presumptions lie behind the preceding analysis. First, it is presumed that the cow, once moved to the lower-cost ration, will continue to produce the same amount of milk she would have made had she stayed on the more expensive ration, or said another way, that she would have stayed on the same lactation curve on the cheaper ration.

If this was true, and a farmer plans to move the cow on Monday, one could ask why not on Sunday? If a dairyman can move a cow from an expensive ration to a cheaper one without affecting production, why not do it sooner? Using this logic, one would push the group move date earlier and earlier in the lactation until either the whole herd is fed the cheaper ration (a mistake) or one is forced to admit that the only time one should move a cow to a cheaper ration is when one expects the production to decline. Presumably, production losses in a moved cow should be worth less than the feed savings.

As expected, the analysis based on average feed cost per cow in the two groups is flawed. The proper analysis should be based on a marginal cow. That is the one that will be moved. Some assumptions are needed for this analysis as well. First, the analysis assumes that the cow, once moved, will continue to eat the same amount of dry matter (DM) as before the move. This may be optimistic, since the social effects and the likely higher NDF of the cheaper ration may drive intakes down, but for this analysis we will remain optimistic about intakes.

Further, it is assumed that any impact on production as a result of the move will be sustained for the rest of the lactation (i.e., that the cow, once shifted to a lower lactation curve, will remain on that curve). Last, it is assumed that milk production will be driven by energy intake and that the new, cheaper ration is adequate in the other nutrients to support the milk production provided for by the energy intake.

If asked, most dairymen will concede that cows decline in production when moved to a cheaper, less energy dense ration. It simply does not pay to try to reduce feed costs at the expense of production. The marginal value of milk is too much greater than the marginal costs of feed.

Most very high-producing dairy herds (more than 26,455 pounds per cow), when asked, will admit that they do not really have a low-group ration. They may feed cows as groups (heifers, breeding cows, etc.), but the driving motive is rarely because they are trying to reduce feed costs.

It is probably safe to say that the overwhelming majority of dairies in North America limit feed to their milking cows. Usually this is done out of ignorance and out of the misguided attempt to reduce average feed costs by reducing the cost of feed fed to a given cow on a given day. Dairymen and professional nutritionists alike focus on the cost of feeds, tweak their linear programs, limit feed delivery or access to the bunk and feed poor or cheap feeds in an effort to reduce average feed costs per hundredweight. The individual dairyman would be well served to feed more and better, even more expensive, feeds per pound of DM, if the resulting milk production increases.

It may very well be that there is reason to group-feed cows in order to reduce costs of protein or expensive additives in later lactation. It is unlikely, however, that it is ever financially wise to deliberately limit the energy intake of a milking dairy cow. The exceptions, such as very fat cows, do not outweigh the enormous losses dairies incur by trying to feed cheaply. Marginal milk is simply too valuable.

30-day dry periods
Every few years, a new fad hits the dairy industry. However, many of these get-rich-quick schemes later turn out to be not so useful. A few examples – extended lactations from delayed breeding and 4X to 6X milking of fresh cows – arrive with great fanfare, but few dairies find them profitable.

Shorter dry periods seem very attractive. For many dairies, they only need one dry cow ration, which simplifies formulation and delivery and makes feeding dry cows much more consistent. In addition, it appears there is small impact on the following lactation. So the opportunity to accrue the additional cash prior to freshening seems like an obvious decision.

This example demonstrates a failure with an improper partial budget. From a single cow perspective, it looks great. But some assumptions were flawed.

First, we assume there is excess parlor capacity and milking cow housing capacity. Oops, stop right there. If there was truly excess capacity in all those areas, a bigger mistake is already in place. Fill the barn! Never run a facility at undercapacity. It is perhaps the most expensive “disease” on a dairy.

So the true comparison is not the dry cow to herself, milking 30 extra days, but the difference between milking a cow eight months pregnant versus milking the average cow in the herd. In every dairy that has tried this, the milk production during that extra month of lactation (the eighth month) is at least 20 kilograms below the herd average. Filling the parlor with these cows is no longer a get-rich-quick scheme.

Sire selection
Sire selection serves as another common example of faulty decision making on dairies. Investing money on better semen pays off in more milk production in the future. That increased milk production will require greater feed consumption. It is important to properly account for the increased costs and increased income to make the correct decision.

Each dairy must decide what price they are willing to pay for semen performance. Semen is always an investment in the future. Even in the hypothetical situation where the semen purchased today results in a successful pregnancy, the gestation period of nine months occurs before the calf is born. In general, there is nearly a two-year growth phase before the first freshening. Finally, there is another year before the first lactation is completed. Thus, the investment in semen is made with current dollars, but the payoff is three to six years in the future.

In addition, the conception rate of the dairy affects the cost-benefit ratio. If the conception rate is two services per pregnancy and half the calves are females, it takes at least four units per female calf. Typically, some inseminated animals are culled, some animals abort, some calves are twins or die, so it often takes nearly five breedings per live calf. Not every heifer gets pregnant, freshens and completes her first lactation to become a second-lactation cow. On most dairies, the annual semen purchases divided by the number of lactations to cows exceeds six straws. If these assumptions are close, then it takes at least six units of semen to achieve the predicted benefit of that sire.

The benefits are best estimated as the predicted difference between two sires. This difference will occur on average for at least two years and will also be partially passed on to future offspring. Thus, a 200-kilogram difference between two bulls will perhaps return 500 kilograms in future milk. Two important adjustments must be made. First, the income will not accrue for three to four years, and secondly the extra milk produced will need extra (marginal) feed.

The goal of breeding a cow is to change the status of that cow from open to pregnant. Money is spent on semen, insemination, heat detection, veterinary examinations and medications. These expenses can be estimated. It is much more difficult to estimate the value of these interventions. What is the economic value of a pregnancy? Knowing the value of a pregnancy allows us to make better decisions about the economics of reproductive interventions, and we can justify the energy and expense in getting a cow pregnant.

Any economic estimates that involve increased milk production need to account for the increased feed costs. These must be calculated using marginal feed costs, not average feed costs. Any drop in milk production is not a complete loss; there is typically some savings in feed costs. However, the value of milk far exceeds the value of marginal feed costs. Getting cows to eat one more bite, thus trying to get them to produce one more kilogram of milk, will almost always produce more profit. It should be a mystery to every nutritionist, veterinarian, feeder, manager and banker, if a dairy ever has an empty feedbunk.

The concept of marginality and marginal decision making is the key to making the right decisions on dairies. In some respects, veterinarians are comfortable in thinking in marginal terms, since each therapeutic decision in an individual cow is a marginal decision in economic terms. Dairy farmers and their advisers can be led astray, however, when they begin to base the next marginal economic decision on the overall historic average performance in a herd or the industry. Management practices that are widespread and well-accepted may not be the best economic choices.

Advisers to dairy producers have a responsibility to think freshly (and marginally) about the standards of management to which we have all become accustomed. Many dairy farmers forego very significant profit opportunities in the false pursuit of reducing the costs of inputs.

By focusing on the costs of inputs and not the inputs’ marginal impact on revenue (milk) and, therefore, profit, many dairy producers box themselves into a cycle of poor investment decisions, poor profitability and a poor lifestyle. Advisers who are properly prepared with the proper conceptual basis can help them break out of that cycle and improve their clients’ livelihoods and lives. PD

References omitted but are available upon request by e-mailing editor@progressivedairy.com.

—Excerpts from 2006 Western Canadian Dairy Seminar Proceeding