As horse owners, we aim to provide the best welfare for our horses, but, what is good horse welfare? And, how can you determine or measure if a domestic horse has good or bad welfare? 

These and many other questions have kept animal welfare scientists busy for many decades. And, last February, at an event hosted by Horse South Australia (Horse SA), we had a unique opportunity to learn the latest in animal welfare assessment from a global leader in the field, Professor Emeritus David Mellor.

Prof. Mellor has served on numerous national and international animal welfare advisory committees and has wide experience of integrating scientific, veterinary, industry, consumer, animal welfare, legal, cultural and ethical interests while developing animal welfare standards, regulations and laws.

Although officially retired, Prof. Mellor has taken a keen interest in sport horse welfare and was invited to assist New Zealand Thoroughbred Racing (NZTR) in developing a new welfare strategy based on the Five Domains Model of Welfare Assessment, a model he developed in collaboration with other scientists, and is used in many areas, from farming to zoos.

I will be reporting on NZTRs welfare strategy when it is released but, in the meantime, I was interested in learning from Prof. Mellor his view on how welfare impacts the sustainability of horse sports, how we can foster healthy welfare discussions within the horse sector, and where the Five Domains Model of Welfare Assessment fits in.

Why discuss horse welfare using the five domains model of welfare assessment?

A brief answer to that question is, “Because you are dealing with sentient animals.”

Sentience is the capacity of an animal to have positive and negative experiences that matter to it and affect its welfare, and the five domains model is a very good vehicle for exploring, understanding, managing, and assessing those experiences.

In New Zealand, every vertebrate is now defined as sentient, but the legislation doesn’t define it in as much detail as the one I just provided above. Europe, NZ, Quebec and a number of other places have declared that vertebrates and some other animals are sentient [and Australia’s ACT is considering doing so soon].

It is my view that specifically acknowledging sentience is a really useful thing, because it acknowledges what animal welfare scientists implicitly and explicitly have been thinking for three or four decades.

While the legislative changes are quite recent and we animal welfare scientists have been doing quite well without them up to this point, I think it is an important step to take.

Especially today, because it formally puts a stake in the ground to say that animals are not merely commodities; they are beings that can often have really negative experiences and, if given the opportunity, can also have really positive experiences.

This legislative recognition of sentience now demands that we keep this in mind when we are going to use animals for our own purposes.

If you take an animal welfare point of view, the principle is that it is acceptable to use animals for our purposes with two provisos;

  • that the purpose is acceptable (which is where the concept of social licence to operate fits in), and,
  • that the ways the animals are used are humane.

And the humane component may be assessed by using the five domains model.

The acceptability by the wider public of the ways we use animals, aligns with the extent to which the animal’s welfare is compromised or enhanced by what we are doing, (and that any compromise that we can’t get rid of is suitably compensated for by what we are doing).

The social licence comes from the recognition of the welfare balance and whether or not that use is regarded as acceptable by the wider public.

Does this mean that welfare is fundamentally linked to the future sustainability of the horse sector?

Absolutely it is, and there are other examples of the way social licence can have an impact and has had an impact and led change.

In my presentation, I gave examples in farming, one of which was about layer hens and the speed of moving to free range layers (the proportion in Britain is now very high). That change was socially led, because the industry did not want to do it initially. But they did it when they understood that people would not buy eggs that were not free range.

But let’s be clear. There’s no way of managing any animal – whether it’s your loved pet or a free range or caged layer without the system used having negatives. This is why in farming we have developed recommendations for ‘best practice’ (although I would much prefer they referred to ‘good practice’), that set welfare at a higher level than minimum standards, which are safeguards to minimise negatives.

The minimum standards are the threshold below which it is not acceptable to go. They also provide guidance to people who are not educated enough (and I don’t mean ignorant, merely that they don’t know essential information), or a basis to penalise those who are wilfully neglectful or even sadistic and, so on.

In terms of improving welfare, do you think that the horse sector can be compared with livestock farming?

In New Zealand, farmers understand that, unless they manage animal welfare in a way that is acceptable to the public in the Northern Hemisphere (where the premium markets are), they will not be able to sell into those markets.

So, what animal welfare scientists, regulators and the veterinary profession did, about 28 years ago, was to press the economic button by emphasising the benefits of access to these markets.

As standards improved, farmers began to see that their animals were in better condition and they began to take more pride in good husbandry, as opposed to production. That distinction, made by John Webster, was a very good one because when you change your focus from animal production to animal husbandry, automatically the emphasis is on care.

In contrast to farming, the orientation of horse racing draws attention to some significant challenges facing the sector. Not the least of these is that such activities are done for financial gain or for human gratification, either by winning or being the person who has the best horse, being the best trainer, etc. Thus, unlike farming, it is not an enterprise that is essential for our survival.

In terms of other equine activities, I don’t think they are yet as vulnerable as horse racing because you can make a case that interacting with animals has been part of our culture for many centuries and it is possible to develop good relationships with horses.

There are areas where it is possible to manage horses in ways that are acceptable, even though there are many ways in which their welfare can be compromised. Like the occasions, hopefully rare, when a person punishes an animal for failing to perform. Such incidents, if observed by the public, would invite close critical scrutiny and could damage social licence.

Do you think the racing sector is better placed, in terms of its social licence to operate, because of its huge contribution to the economy?

That’s a bold statement to make. At the moment it may be true, but I, nevertheless, think the horse racing sector has to give a lot of attention to the way they manage their animals. New Zealand Thoroughbred Racing (NZTR) is making very good strides in that direction, but nothing is impossible, and you can unexpectedly reach a tipping point in social licence.

This is what happened with, for example, the challenge against research, teaching and testing using animals by animal liberationists in the UK. They challenged again and again with increasing assertiveness. But a negative tipping point occurred when the activists desecrated graves of people they opposed. This so appalled and incensed the public that the Prime Minister and other politicians made statements that research, teaching and testing was supported by the government provided that it was done according to the rule of the law. The government had never made such an explicit statement before.

As another example, if it were shown that you could actually race horses bit-free and they performed as well, were controllable and safe to ride and, especially if it were also found that they performed better due to a clearer airway; and, if it were simultaneously shown that surgical procedures deployed to correct quite common palatal and laryngeal infirmities were required far less or not at all; that might lead to a positive tipping point with regard to bit-free riding.

What do you hope the five domains model can achieve for the horse sector?

My hope for the five domains model is that users will appreciate its flexibility and will adapt it to best suit the species of their interest, in this case the horse. I am not a five domains evangelist. The model is not a definition of animal welfare. Rather, our latest understanding of animal welfare underlies its primary use, which is an instrument for thoroughly assessing animal welfare in order to guide its management.

So, my purpose here is to say that our understanding of animal welfare evolves. I have been explaining what our latest thinking is about animal welfare, and, importantly, that this thinking is going to evolve overtime. Of course, at each stage we will only incorporate the thinking that is well validated.

If you are developing a welfare strategy – and I see there is an impetus to do that now in the sport horse sector – and if you are educating people about the welfare of horses or any other animal for that matter, the five domains model is a way of assisting that process. There’s a lot of work to do, but what NZTR has taken on board is good place to start.

When I became involved, I didn’t go there to tell them the things that are wrong (which is what most people want to do); rather I said, let us examine all the things you are doing well and how you judge that.

We worked through that using the five domains model as a team with two veterinarians, a trainer and Martin Burns; we laid out all the positives and we changed elements of the five domains model to be more suited to what you could say about horses.

When you can recognise what is going well, automatically, you start to think about what happens when things go wrong, and you have to think about how you would recognise this.

What NZTR have done is to say: Here’re all the things we have in place to help ensure that a horse is well fed, in a good environment, is healthy, has opportunities for behavioural expression and to have positive experiences.

Then it becomes easier to see that horses which have particular characteristics will fall short in various ways and are in situations that do not meet minimum standards. And NZTR did that spontaneously.

What I try to do is to rigorously show how the model can be used; in my presentation you saw seven applications.

I want to show that it is equally, if not more important, to identify what you are doing right, because if you can do that, you can automatically set a standard for when things are not going as well or really badly. It’s something people naturally think about without being strongly aware of doing so.

The five domains model is not set in concrete. It has been adapted by NZTR for racehorses, it’s been changed for whales and tomorrow I’m having a chat with Dr Andrew McLean to see if we can make changes that are appropriate for elephants. I am also looking at it in regard to snakes, which is tricky because, how do you know when a snake is smiling?

Of course, if you were dealing with very different species (and horses and farm animals are not very different), such as exotic species that have sensory capabilities that we don’t have, use of the model would involve a high degree of speculation.

Nevertheless, its cautious use can still be informative, even if only to draw attention to areas where we need more information.

A good example is cetacean senses like echo-location. I don’t know what the experience of echo location is like. But we have knowledge of cetaceans’ natural oceanic habitat in which they migrate for thousands of kilometres and encounter a huge diversity of environmental circumstances with which they engage.

Thus, it is reasonable to assume that keeping them in a concrete bath, no matter how big the concrete bath is, would be like putting them in a prison. With regard to echo location, the restricting concrete base and walls would likely create a reverberating container, which it does seem reasonable to assume would likely be very unpleasant.

These and other reasons revealed by the five domains model lead to an inevitable conclusion that keeping cetaceans in captivity is unacceptable, a conclusion which is progressively leading to a loss of social licence for this practice to continue internationally.

So, what do we do next?

Returning to the sport horse sector, becoming familiar with our current understanding of animal welfare is a first step. A second step is to use the five domains model to thoroughly evaluate the welfare status of sport horses in the various ways they are managed before, during and after events, and, in the longer term, as they become less competitive and age.

This involves understanding what needs to be done to achieve good welfare management, how to recognise different levels of welfare and how to correct or avoid welfare which is not meeting acceptable standards.

A third step is to outline a strategy for improving sport horse welfare, where that is required. This would include details of specific welfare priorities and targeted dates for completion.

A commitment to improvement and demonstrating by action that stated targets are being achieved, would help to retain social licence more successfully than “all talk and no action” will.


  1. Become familiar with the current, evidence-based understanding of animal and horse welfare.
  2. Use the five domains model to thoroughly evaluate the welfare status of sport horses in the ways they are managed during their different life stages and for different activities.
  3. Outline a strategy for improving sport horse welfare where required, setting specific welfare priorities and targeted dates for completion.
Horse Welfare Assessment Guide

Click on the image to download your free Horse Welfare Assessment Guide.

Horse Welfare, the Five Domains Model