Replacement Strategies to Maximise Profitability in Dairying

3.1 Introduction

The objective of this section of the study was to update the information available on dairy cow disposal rates on commercial dairy farms. The data used were assembled from the DairyMIS computerised recording system and from the National Farm Survey (NFS). The farms included in the DairyMIS system are representative of intensive dairy farms and are mainly located in the South of Ireland. The NFS database is representative of dairy farms nationally.

3.2 Results and Discussion

The disposal rates of cows in the NFS database for the years 1990-93 is shown in Table 1, with categories for dairy herd size.

Table 1 : Cow disposal rate by herd size category (NFS data)

The average disposal rate was 15.6%, ranging from 12.9% in 1990 to 17.9% in 1992. The range in cow disposal rates for different herd size categories was also relatively small (14.3% - 16.2%). The dairy herds participating in the DairyMIS system are representative of the larger more intensive dairy farmers in Ireland.

The average culling rate from the DairyMIS data set was 15.2% compared to an average culling rate of 15.6% from the NFS data base. The cow disposal rates for farms in the different herd size categories was relatively similar. This culling rate is lower than that found in a similar study in the early 1980s (Crosse and O'Donovan, 1989), when the average culling rate was 17.6%.

The disposal rates of cows by primary reason for disposal in the DairyMIS database are shown in Table 2. Up to 31 primary reasons for culling were recorded initially but it was found that five reasons consistently accounted for over 60% of the total disposal rate. The remaining group of 26 less important reasons for disposal were therefore grouped together under 'other reasons' for this analysis.

Table 2: Cow disposals (%) - by primary reason for disposal (DairyMis data)

Chi square value = 80.6 ***

The most significant primary reasons for culling were infertility (3.6%), surplus (2.2%), mastitis (1.8%), and low production and old age (2.0 %). While there were differences in these figures between years, the relative importance of these reasons remained constant over the years. The major advantages of culling strategies for herd improvement are achieved by maximising the proportion of cows that are culled for voluntary as distinct for involuntary reasons.

Generally, culling for low production is considered to be voluntary culling. Since the advent of milk quotas, culling of surplus cows has assumed much more importance, and it is likely that dairy farmers would cull their poorer animals if they had surplus stock. Culling of surplus stock together with culling for low production is therefore considered to be voluntary culling.

In this study, it was found that about three quarters of all cullings were involuntary and only one quarter were culled for voluntary reasons. Infertility was the most important single reason given for involuntary culling and while culling for infertility was particularly associated with older cows, a significant proportion of younger animals were also culled for this reason. Mastitis was the next single most important reason given for involuntary culling and this was associated more with older cows.

The disposal rate of cows was significantly (P<0.001) influenced by the parity (lactation number) of the animal (Table 3). The highest incidence of culling (20.8%) within each parity category was for cows with greater than 4 lactation's. This was associated with infertility, mastitis, age, low production as well as a collection of other reasons. The culling rates for the other parity classes were relatively similar. The relative ranking of animals within different parities in terms of disposal is important. The potential genetic gain from a selection policy is maximised by the disposal of older cows of inferior genetic quality.

Table 3 : Cow disposals (%) - by parity (lactation number) of animal (DairyMIS data)

Chi-square value = 27.8 *

The reasons for culling by parity group are shown in Table 4.

Table 4: Cow disposal rate - reason for disposal by parity of cow (DairyMIS data)

Chi square value = 149.9 ***

Infertility was the most important main reason for disposal, though for first lactation animals it was ranked second. A surplus is the second most important reason accounting for 14% of all cullings, though it assumes a somewhat higher importance in first and second lactation disposals. Mastitis, which accounts for 12% of cullings overall, was particularly important in later lactations.

Low production, on the other hand, was a relatively important reason in early lactations, but not later. The remainder of cullings were attributed to a very large number of reasons each of which accounted for only a very small percentage of cullings, and have been grouped together under "other reasons" (Table 4). While each reason was only of minor importance, the grouping accounted for approximately one third of all cullings.

The seasonal distribution of cow disposals for the years 1990-1994 is shown in Table 5, and it can be seen that there has been year-to-year fluctuations in the pattern.

Table 5 : Cow disposals (%) - by period of year (DairyMIS data)

A quarterly breakdown of the data shows that the period April to June accounts for the highest percentage of annual disposals averaged over the five years, but it does not represent the peak in every year. In general, however, it can be seen that the period January to June accounts for a high proportion of cullings, typically two thirds. During the other six months, July to December, disposals are typically low, though a quarterly breakdown shows occasional peaks for particular years. While most of the herds recorded were spring-calving, some of the distribution of cullings throughout the year may be due to the presence of some Autumn-calving herds in the survey.

Analysis of a matched sample of 23 farms showed that they had an average disposal rate ranging from 10% to 20%. The variation in disposal rates both across years and between individual farms was significant. The large farm-tofarm variation in culling rates recorded in this study as in the previous study would indicate that management is important. A reduction of involuntary culling should allow for more culling for low production and other reasons, which may enhance the genetic quality and general health of the dairy herd.

4.1 Development of a simulation model

The breeding goal on many commercial dairy farms is to maximise profitability, and in this regard, the genetic quality of the dairy herd is an important determinant of farm profitability. There has been a marked increase in the rate of genetic improvement in dairy herds in Ireland in recent years. These developments present new opportunities for increased profit on dairy farms as well as major challenges to milk production systems that are mainly based on grass.

The simulation techniques known as Monte Carlo methods were used in this study. A model using Monte Carlo methods was developed to simulate the physical and financial consequences of alternate breeding goals and alternative levels of voluntary culling for a hypothetical Irish spring-calving herd. The characteristics for this herd are based on the Moorepark Blueprint for summer milk production. This model farm represents a typical intensive dairy farming system in Ireland where the EU milk quota is the most limiting constraint on production and where there has to be a combination of enterprises to use the land available. This allowed the opportunity cost of capital and land to be evaluated.

4.2. State variables included in the model

A cow's relative breeding index (RBI'95) describes, in a single figure, a cow's genetic merit for the production of fat and protein. It also includes an economic component. RBI'95 was included as a state variable for each animal in the herd, as were lactation number and production level. Production level was described by one of 15 levels and defined relative to the mature cow equivalent production. In this model, the lactation of an animal was measured as 1st, 2nd and 3+.

4.3 Other Model Parameters

Current input and output prices were used in the analysis. An involuntary culling rate of 18% was assumed. It was further assumed that 50% of calves born were male and that an overall calf mortality rate of 6% prevailed. The RBI of a calf was calculated as the average of the sire and dam's RBI'95. At the end of each year, new heifers entered the herd until the quota could be expected to be met in the following year (although production levels for the following year were unknown, expected yields on the basis of current production levels could be calculated). Replacement heifers are 2 year-olds, which were bred and reared within the herd. The RBI'95 of these heifers was assumed to be equal to the average RBI'95 of female calves surviving from 2 years previously. Current (1997) costs and prices were used for milk, calves, cull cows, replacement heifers, non-feed costs per cows and for the opportunity cost of land and capital.

The RBI'95 of an animal in the herd had an effect on physical performance and on financial factors. The following were dependent on the RBI'95 of animals in the herd; milk yield; feed costs; calf value; carcass value; land usage.

4.4 Culling and Breeding Goals Tested by the Model

The physical and financial performance factors were simulated for two breeding goals: (1) where sires with an average RBI'95 of 130 were used: (2) where the average RBI of the sires used was 150. For each of these breeding goals, the RBI95 of the average sire was allowed to increase (by 2 units per time stage) over time to reflect the improvements in sires used over time (i.e. RBI Sire (t) = BRI sire (0) + (2*t).

Three contrasting levels of voluntary culling 0, 10 and 20% were also evaluated. At the end of each stage of the simulation, a number of animals (based on the current herd size and on the voluntary culling rate) were culled according to the lowest lactation adjusted milk yield). At each stage of the simulation, the total yield, total revenue, land in use, and average RBI of the herd were calculated.

4.5 Stochastic Elements included in the model

A planning horizon of 20 years (20 stages) was used. At each stage, each animal in the herd could be culled involuntarily with a probability of 18%. Survival was determined stochastically, as were the transitions from one production level to another. Transitions in production levels were assumed to take placed at the end of lactation (stage).

For each calf in the herd, sex and survival were also determined stochastically. Transitions in 'lactation number' could be determined deterministically since only one transition was possible for an animal at any stage (i.e. lact (n+1). An animal's RBI'95 remained constant for its lifetime, so again the transitions for this stage variable were deterministic.

4.6 Results of physical and Financial Factors

As the simulation model was run, certain physical and financial factors were measured at each stage. These were; Average RBI'95 of the herd; Total milk yield of the herd; Average milk yield per cow; Current herd size; Land in use for dairy enterprise; Capital investment; Total revenues.

Breeding the herd to high index sires was found to have large effect on the average RBI of the herd

For the later half of the planning horizon, a higher breeding goal resulted in higher herd RBI95's, no matter what culling rate was chosen. Where the lower breeding goal was combined with high voluntary culling rates (20%), the average RBI95 of the herd rose quickly, but within 6-7 years, the average RBI95 where the higher breeding goal was used (regardless of the culling rate) was always higher. When there were no voluntary culling, increases of 18 and 35 units of RBI95 for years 10 and 20 respectively, were recorded for the lower breeding goal (sires with RBI95 130).

This compares with an increase of 24 and 47 units of RBI95 respectively, for the higher breeding goal (Sires with RBI95 of 150). The increases in RBI95 associated with higher culling rates are due to the faster introduction of genetically superior animals to the herd. The increases due to culling policy are greater for the higher breeding goal because, with the lower breeding goal, the genetic level of the herd soon approached that of the replacement heifers. The effect of breeding and culling strategy on milk yield is given in Figure 2.

In the model, a herd member's milk yield was adjusted for RBI'95 level, So, for a certain age and production level, an animal with higher RBI'95 would have a higher yield than one with a lower RBI'95 level. Therefore, it is not unexpected that the use of the higher breeding goal (which results in higher average RBI'95 per animal (Figure 1) results in a higher milk yield per animal as shown in Figure 2. Because of the inclusion of a milk quota higher milk yields per cow (associated with higher breeding goals) meant that herd sizes had to be reduced, as illustrated in Figure 3. When a voluntary culling rate of 10% was applied, herd sizes of 85 and 76 were recorded for years 10 and 20, respectively, when the lower breeding goal was applied. While for the higher breeding goal rather than the lower, results in a 11% reduction in herd size after 10 years and a 30% reduction in herd size after 20 years.

The land used for the dairy enterprise was based on the genetic makeup of the herd. For the higher breeding goal, the average RBI was greatest (Figure 1). However, the fewer animals required in the herd to reach quota (Figure 3) had the effect of reducing the land area required for the dairy enterprise (Figure 4). A measure of farm profit was taken at each stage under the various breeding and culling strategies. These measures are shown in Table 6. The calculation of these figures was based on milk revenue, calf sales, feed costs, replacement costs, land usage and capital expenditure.

A cull policy with no voluntary culling resulted in higher profits (Table 6) than either a 10% or 20% voluntary culling rate (and the 10% rate resulted in higher profits than the 20% rate). The use of higher merit sires (150 RBI95) resulted in higher profit rates over the 20 year planning horizon.

Table 6: Change in Actual revenues (per year) relative to a breeding goal with 130 RBI sires and no voluntary culling (set at 100%)

4.7 Conclusions from the Simulation Model

The results of the simulation suggest that the use of high merit sires had a large effect on the genetic merit of the herd and consequently on milk yield per cow. This resulted in increased farm profit even though cow numbers had to be reduced to stay within the quota. Increasing the rate of voluntary culling also increased the genetic merit of the herd, especially in the early years. This had a positive effect on milk yield per cow as well as on other farm parameters. However, the cost of achieving this higher milk yield per cow was outweighed by the corresponding higher cost of providing replacement animals. Therefore, culling on the basis of production alone resulted in a reduction in total farm profit.