Saturday, March 18, 2017

Genetic Value in dairy within a matured genetics industry

Greg Palen


Do we see the market as it is or do we wish the market to see us differently than it does?
                Semen pricing realities

No one active within the dairy AI industry is going to have the full perspective of the market as it really exists in total.   Broad statistics will suggest that the industry is serving a matured market - the industry has done its job in optimizing its penetration with existing domestic customers but is making no headway with those not yet (or no longer) using AI.   Thus any growth in sales comes from foreign markets that only recently opened up to foreign trade, or from selling in advance of consumption.   With no change in the delivery technique for dairy genetics (cervical entry, rectal palpation guiding semen placement) or in the technology to keep germ plasm viable (cryogenic storage containers requiring periodic liquid nitrogen recharge) the basic AI technology is seen as fully matured.    Lack of industry emphasis on technique training and conception monitoring has meant the remaining non-users are not converted, thus promotional efforts based upon “genetic value” (neglecting its management value) merely reinforce status quo levels of dairy investment in the use of AI and tend to merely reinforce existing market shares.

The statistical evidence actually suggests that the majority of the market (dairy farmers using AI) perceive the product of the industry (semen) as a commodity.    Economists will tell you that you are marketing a “commodity” product when it is difficult to dislodge established market shares on any basis other than price concessions.   Relative disinterest in the marginal differences in product lines’ “genetic value” suggests the latest generation of dairymen, while college trained and thus more responsive to new technologies, are perceiving AI and genetic indexing as “old technology”.     Thus they do not allocate as much money toward marginal genetic value as they do to feeding equipment, milking equipment and manure management technologies.    There is more willingness to pay extra for genetic diversity and yield advances in forage and grain seeds as this can produce a more immediate profit gain than comes from generational changeover in the production cow herd.

The only aspects of AI that truly are “new” technologies are (a) Ov Synch protocols applied on a herdwise (rather than a therapeutic) basis, (b) sexed semen options, and (c) daily heat detection by pedometers linked to computerized production monitoring systems.     These three areas have helped to maintain use of the AI technology, which was really in danger of declining within large “expansion” herds in which one of the goals is a reduction of labor overhead per cow.   

In fact, I think a strong argument could be made that the only impact from “Genomics” (which is perceived as a revolutionary technology, yet in practice is merely an evolution of the established  reductionist methods for indexing “genetic value” in additive performance traits) has been that it has moved the attention of commercial dairyman away from the rapidly growing AI practice of dairy crossbreeding.     “Crossbred” is now the second most populous breed behind the Holstein, ahead of the equally fast growing Jersey breed penetration of commercial dairy operations.   But in both cases, as “XB” or “JE” replaces “HO” in so many operations, there is an undercurrent in this that is rejecting the basic philosophy of “genetic value” as necessary for dairy management to improve profitability by increasing commodity milk production from “genetic value”.

(page two)

To summarize, the perception that “genetic value” drives the bus - and thus is the source of all premium valuation for high genetic index rank animals, is being marginalized by the dairyman emphasis on creating a synergistic dairy operating unit.    Dairy AI has become very insular in its focus on the genetic value “horse race” and is not listening to feedback from its actual customers that is expressed in their pseudo-loyal brand buying behavior, lack of response to advertising, and unwillingness to increase the per straw price they will pay for semen that the industry has produced.   

After six years of Genomic sire offerings, has the market bought the technology?

First, let us actually admit that when it comes to Genomics, the dairyman has no choice.   There is no established AI system buying bulls on any other basis.    To access a “non genomic” sire, ie, a sire selected on any basis other than this most reductionist of all genetic approaches, you have to collect him yourself.     [ Note:  NAAB reports from CSS affiliated custom collection centers show double digit annual growth in units of custom collected dairy semen for the recent decade.   Genomics does not appear to have changed this trend.   In each herd wherein Genomics is perceived to have failed to produce good cows, this trend will accelerate. ]

The majority of dairymen are not interested in genetics to the level of doing their own selection, thus they accept what the AI industry offers them.    Their general approach is to follow “matrix selection” [in spite of decades of experts advising them to select on ranking indexes alone] thus they specify minimum combined levels of milk, fat (or fat%) and protein (or protein %), type as an emphasis on Udder and Foot & Leg composites, and more recently, health/fitness traits (SCC or PL and/or DPR).    Much of this will be determined by the milk market payment system they are in.   Competing AI salespeople will present the sires that meet or beat criteria and then prices are negotiated.   He who offers the best prices “wins” the sale.

In fact, those who select from index rankings are a minority comprised of purebred breeders who wish to compete in High-Genomic propogation and marketing, and those commercial dairymen who have less interest in genetics than the “matrix” selectors and thus just follow advice from some dairy management consultant (who is probably more interested in ration balancing).   

These dairymen tend to avoid actually milking their own cows and thus are not engaged in all the questions as to whether genetic selection and/or mating selection could produce new cows more “adaptable” to the cow environment and production process dictated by their facility designs.
If there is a design flaw in reductionist genetic value theory, it is the assumption of a “model” environment, thus an accumulated statistical ignorance of how differences in environments will affect the ability of the dairyman to “harvest” the genetic “value” they are utilizing.

In the 1970s-1980s, as the previous generation of dairymen transitioned from pedigree to index selection of mating sires, “genetic value” earned a premium in the cow market for anyone willing to utilize AI.    Thus, there was clear differentiation between perceived sire rank and his realized sales price ($5.00 for basic bulls, $10.00 for good bulls, $15.00 for premium bulls, $20 to $50 for “sires of sons”).      Purebred breeders competed for the best “progeny proven” sire semen.
 (page three)

Today there is no detectible genetic response in cow prices.    Any volume auctioneer or treaty seller of dairy cattle will tell you the biggest determinant of premium cow sales values are:  (1)  “fresh” stage of lactation, (2) younger age, first or second lactation (3) visual evidence of well-grown and healthy, (3a) clear premium for being “typy” in udder, frame, feet and legs; (4) prior reputation of the seller.    The floor value is set by beef salvage recovery at the time.   

There is really very minor movement of “pedigreed” cattle, wherein “index rank” may become a factor in price determination.     Supporting this is how often today, when groups of siblings who result from ET propogation of a Genomic based mating are offered, the one who has the highest imputed value sells for five figure prices—all the siblings bring four figure prices, some truly at the commercial market price level.    The market for high Genomics is thin because there is no one to buy their offspring other than an AI system with a MOET program or a breeder wishing to compete in that market, and a perception that the sales opportunity depends on IVF propogation and will be gone by the time the animal reaches “normal” lactation production.    

This “value” does not transmit to the commercial dairyman who is not pursuing genetic cattle sales.     Commercial dairymen purchase 95% of all semen produced.    The floor price for the semen they buy is established by competitive behavior among the AI marketers in their region.   The marginal premiums they will pay depend on their perceived ability to recover them, as a result of various factors:
(a)    Conception rates normally achieved
(b)    Success of calf raising and heifer reproduction program
(c)    Expected herdlife of each replacement entering the herd
(d)    Opportunity to market surplus cows, especially above costs of raising heifers

The promoters of Genomics have uniformly failed to establish how this technology benefits a dairyman.    It clearly benefits the index-marketing AI stud,  because it helps you avoid buying bulls who have lower probabilities of ranking high as progeny-tested sires.   However, the irony is that “unproven” young Genomic sires are replacing “proven” sires validated in the market by realized performance.     Forcing the validated sire to compete directly in the market against the theoretically imputed sire on the same genetic scale assumes an acceleration of genetic ability that the market has a long history in disproving.

The decline of purebred breed association activity, and the usurpation of traditional purebred methods for realizing cow value into the “sire stack” pedigree valuation on which indexes are based, has reduced the number of people benefiting from (d) as it relates to “genetics” even as it increases the number of people benefitting from (d) as it relates to use of newer technology such as sexed semen or improved health and fertility (producing more heifers per herd unit).

As for the purebred sector, there remains the market niche which values animals as much for the maternal line development as it does for the (increasingly similar) AI sire stack.   This will likely be the only salvation route for purebred associations, who must identify a way separate from the sire-based Genomic imputation, to produce a cow-based standard for premium market valuation.

(page four)

To summarize, marginal genetic index value has a declining effect on dairy herd equity and surplus cattle cash flow.     The majority of cattle sales are for herd expansion or replacement.

 As a result, semen prices no longer differentiate much, because most dairyman are not planning to harvest incremental sale value;  and the more hungry AI sales systems (driven by the market’s clear perception that there is a “surplus” of high genetic value semen) have created a perpetual price war at the farm gate to substitute for the growing inability to distinguish one sire program “brand” from another.

What do dairymen currently pay for the majority of semen used?

In the Michigan market, as an example of a “top ten” dairy production state, there are under 3000 dairy herds milking slightly over 300,000 cows (average 110 cows per herd).     100 dairies (at a size of 1000 to 5000 cows per location) produce over 50% of the Michigan dairy production.

All major AI systems—Select Sires, CRI Genex, ABS Global, Semex USA, Accelerated, Sexing Technologies (Taurus division), Alta Genetics, plus CRV Holland Genetics and LIC New Zealand– have employee or commission contract sales personnel based in Michigan.     There are in addition five major independent semen sales individuals or companies, buying at wholesale as resellers.   ABS and Select Sires, to a lesser extent Alta Genetics and CRI, have major “captive” insemination systems that have targeted larger herds, thus may have monopoly control over a third of Michigan semen sales.   

ABS’ large herd AI tech program offers semen in the cow at $14 per service, no matter the list price of the bull.    Their GMS computer mating program determines which sires are used.   If the dairyman buys “outside” semen it is put in the “third choice” column and rarely gets a cow bred.
Select’s large herd AI tech program charges $2 extra per service for use of “outside” semen.   In the Alta Advantage program, you basically get 50% off everything but must use 85% Alta semen and half of that usage will be “sampling” sires (Genomic selections).
  
Thus you have 300,000 dairy cows and 120,000 heifers to breed annually, 75% are bred AI thus 315,000 animals to AI, with an average of 2.5 straws used per animal (3x per cow, 2x per heifer) or a market potential of 787.500 straws annually.    One third of that is captive to established AI systems utilizing inseminators and/or computer mating programs.    That leaves 525.000 straws of semen sold among a dozen competing marketing systems utilizing a total of 50 sales people at the farm gate, or an average of 21,000 straws per salesperson per year.    The actual range may be 5,000 to 50,000 straws per salesperson annually (part time independent vs full time company subsidized employee).      

The average selling price of semen in Michigan is likely $12 per straw.     If 50% of the average selling price is allocated to the salesperson, this leaves $6 per straw for the wholesale recovery at the AI stud.     I know of individual dairies as small as 400 cows paying $8 per straw in exchange for an exclusive arrangement giving 100% of sales to Accelerated Genetics.      I know of a 1400 cow dairy that puts his annual semen requirements out for bid, and got $8.65 from Select Sires.
(page five)

What does it cost an independent to sell semen in Michigan?

Suppose you sold 20,000 straws annually and had a $6 average margin competing around this $12 selling average.     You would have $120,000 in gross margin, spread over 50 weeks that is $2400 per week cashflow margin.      You drive 1000 miles per week, get 15 mpg driving a van or pickup with a bulk nitrogen tank (without nitrogen you cannot gain exclusive customers).
You eat $20 daily and spend a couple nights per week in $60 motels.

You will end up with $150 per day out ($600/week) in gas/oil/tires, motel and meals.
You will have $150 per week in nitrogen acquisition expenses (filling your own and farm tanks)
You will have $500/month in payments on truck/van and any equipment utilized.
You will have $100/month in cell phone and other communication expenses.
You will have $150/month business (vehicle, liability, inventory) insurance.
You may have $250/month interest costs to finance inventory and customer charges.

Thus over a year, you spend $49500 on the above.   This leaves you $70500 in net margins, to cover these additional costs:
Any advertising you do locally, any 4H fair or breed club donations.
Any supplies needed for invoicing, postage for statements, promo flyers, etc.
Any shipping costs related to semen and AI supply inventories.
Any cost for garaging and warehousing, and/or office space.

The point is, you could be making $50,000 a year, but it would take 400 straws of sales per week to do that.   You would pay your own health insurance, beyond typical living expenses employed people experience.    You will not get rich, but if successful it will be a steady living.

As an independent in Michigan you will need a base of 200 farms (29500 cows and heifers) and capture one fourth their business (18,500 straws) to produce this level of sales.     It is typical of
dairymen not captive to a single AI organization to buy half from company A, then divide the other half among two other salespeople (keeps the main company “honest”).     Perhaps you will end up that key supplier in a quarter of those herds you visit regularly.

To average $12 per straw you will sell the bulk of your semen in a range from $ 8 to $20.    You could approximate this as:  $16-$20 first service, $10-$14 second service, $ 7-$8 third service.
All of these prices reflect conventional semen—if you can get half of your heifer sales converted to sexed semen, and the margins allow it, you may be able to fund a retirement.

Unless the salespeople make an adequate living and it does not require over 60 hours of weekly travel to accomplish the sales, the AI system will have to deal with a steady turnover of reps.
Family pressures build and unrealistic sales goals burn out the salespeople in many AI studs.
The low level of purebred activity with few breeders profiting from Genomic marketing means 85% of sales must fall within the competitive price range to keep the sales rep employed.    If there is not an established base of business, the AI stud must subsidize initial sales activity or else submit to the unproductive price wars others are using to “buy” market shares.
Is the entire market now driven by Genomics?

At this point, after six years, Genomic sires breed 45% of the cows in the Holstein and Jersey breeds, maybe 25% in Brown Swiss.    Genomic penetration has plateaued.

This means 55% of the cows are still being bred (by preference) to progeny evaluated sires.  
For this market share, the most aggressive pedigrees (young sire, unfreshened dam) have no appeal and expectations of marketability to this sector require the pedigree to “fill out” with realized phenotypic milk and type data to support premium marketing after progeny data is added to the Genomic data in their evaluations.

Reflecting this, you still see major AI systems—Select Sires (largest in USA) and Semex (the corrolary in Canada) – giving preferential presentation to high Rel% progeny-tested sires, and while actively marketing Genomic sires, still presenting them as “young/future” sires.

What does 45% penetration really mean, given 100% of new AI sires enter studs based on a Genomic imputation?

It means that the third of AI herds that are “captive” under contract insemination or computer mating services are not concerned over whether the stud is giving them G or PE sires.

It means that AI studs who force out progeny tested sires in favor of higher numbers on G sires are forcing higher Genomic sire use in their “loyal” herds (eg, CRI Genex).

Other bases of “value” reflecting various established (aAa) and emerging (polled) market niches will be an equal aid to any AI stud wishing to increase market penetration.

After six years of use, the market is getting more sophisticated, recognizing that the “top 1%” of Genomic imputed sires may be the most overestimated.    Thus the use of top 1% G sires at high prices will not grow beyond its current level of gambler breeder and AI mating utilization.    The AI stud that accepts more pedigree variety and development and works harder at providing aAa variety will find marketable sires below the 99 percentile level of Genomic value.

Additive genes are only one third of the total gene selection equation


 Quantitative (additive) genes-- that can be measured on a linear scale intergenerationally
                     (genetic evaluation)  traits that accumulate from specific selection
Qualitative (characteristic) genes—qualities that can be observed intragenerationally
                    (physical analyzation)  cobalancing between physique and performance
Interactive (eg, Epigenic) genes—traits and qualities that switch on or off                        .
                    (the result of environmental triggers) 

The best sire program would account for all three of the above, rather than focusing only on the additive gene competition of index ranking.    Dairymen need more than genetic potential, they need the ability to make synergistic matings that produce physique adaptability, allowing genetic potential to be expressed in the phenotype.    Index by itself is not a tangible “value”.

No comments:

Post a Comment