In a recent video we’ve chatted about why cold and flu medication may be cheaper in winter than in summer – something that may seem a bit counter-intuitive. This was an issue discussed back in 2008 here and here.
For those who aren’t keen on listening to videos, I’ve popped the transcript just below 😉
Cheaper medication in winter
I’ve been doing a bit of online shopping. As part of this, I’ve been struggling to figure out what I should stock up on before the next COVID variant hits.
When looking at the price of Nurofen I noticed that there were only a few specials among the varying brands, and the overall discounts were quite small compared to the deals I’m used to seeing during winter.
Good point Gulnara, this is an observation I’ve made in the past, implying that this isn’t just due to current circumstances.
However, this feels a bit weird – why would cold and flu medication be cheaper in winter than it is in summer? Let’s formalise why this feels weird, and then see if we can provide an explanation.
Demand and price
To formalise our intuition – and start trying to work out why the observed data looks so different – it is useful to build a model. And the best model to start with is our friend supply and demand.
In our discussion of income and price effects we mentioned the idea of a demand curve.
This demand curve tells us that quantity that will be demanded by consumers at different prices.
Here we are looking at the market demand curve for cold and flu medication. As we move down and to the right of the demand curve, additional consumers become willing to buy a product – or existing consumers are willing to buy more. At each price that additional purchase refers to our marginal customer.
Start with the situation where no provider of cold and flu medication is large enough to influence the price with its decision to produce. In that case we get a market supply curve that represents the marginal cost of production of cold and flu medication.
This marginal cost is shown as increasing in the quantity supplied – why? Well the more that is sold, the more storage and shelf space is required, the more need their may be for staff to work overtime, and the more risk there is that the machines creating and packing the medication will break down. The more “scalable” the process is the flatter this curve is likely to be – but the limitations of storage and shelf-space – what are called capacity constraints – point to an increasing opportunity cost from selling more, which makes this type of curve believable.
As it is a competitive market, the price is set where marginal cost equals the price the marginal customer is willing to pay, which is where the supply and demand curves intersect. Even if competition breaks down somewhat (i.e. monopoly or monopolistic competition) the argument given here still holds – so it isn’t too extreme an assumption.
Cool, so what?
Well we want to think about what “summer” and “winter” are here? Imagine a world where you only buy cold and flu medication when you are sick – this would be the same world I live in. Well, I tend to get a cold or flu much more often in winter than summer.
As a result, at a given price I would buy more cold and flu medication in winter than in summer. We can represent that with a demand curve that shifts to the right in winter.
The names for this are “high demand” and “low demand” states – where winter is a high demand state and summer is a low demand state. A high demand state leads to the demand curve shifting right – as for a given price the quantity demanded is higher.
Now here the new market price will be higher due to the demand curve shifting right. This matches our intuition, but not our data!
Gulnara, what is going on?
Good question Matt. To my mind there are six different arguments against this logic. Let me tell you about five of them, and save the last one as a treat for later.
Firstly, COVID has changed things – and so even though it is summer perhaps it is a high demand time.
Secondly, perhaps cold and flu medication is a durable good so prices do not change much over the year as lower prices eat into future demand.
Thirdly, we may have confused average and marginal benefit in this discussion – it may be that the willingness to pay of the marginal consumer is higher in summer than in winter, perhaps because the weather is better so people want to push away the symptoms in order to enjoy the outdoors. More broadly, summer time customers may be “less sensitive” to price than winter customers.
Fourth (and related to the third), as demand for medication is higher in winter, supermarkets may use it as a loss-leader to attract customers to buy other products.
Finally, such products are produced at scale and so the average cost of production is higher during summer.
Each of these five arguments are good, but have a fatal flaw.
The first of these arguments may be true of this year, but as this is an observation Matt’s made repeatedly over his long life, I don’t think it quite covers this.
The second argument would make sense if the price never changes. And the ticket price itself is fixed for these products, so that sounds good. However, it appears that the discount price is lower in winter than in summer.
[show prezi 2]
The third argument is neat, and when combined with imperfect competition it is very compelling, but from personal introspection I don’t know if I’m convinced. Irrespective of the weather I have a given sickness threshold where I will have medication. This may well be part of the pricing behaviour, but we’re going to move on and look for something else for now.
The fourth argument is quite attractive in some sense, but there is something unintuitive about imagining supermarkets cutting prices when demand for things are high in order to try to sell other products – there needs to be something else here, either about the elasticity of demand as discussed above, the cross-elasticity of demand, or with regards to competition as we will touch on below.
The final argument sounds compelling, however the focus on average cost is misleading. The “costs of scale” are fixed costs that must be met irrespective of the amount produced, and as a result it is the marginal cost in the short term – which will either be flat or rising with output – that should determine price.
Getting a bit fancier
Moving away from pure ECON101 we might also note that there isn’t necessarily a competitive provision of cold and flu medication in New Zealand. We only have two supermarket chains which face weak competition from pharmacies for these products due to halo effects.
Halo effects are the idea that if a firm is offering one product people are willing to buy, then people will be more willing to buy other products there – either due to decreased costs of searching as things are all in one place, or because people trust this business to provide a good product.
Supermarkets bundle a lot of goods and services together and so these halo effects are important for understanding the pricing behaviour of these types of firms.
Furthermore, we can look further up the supply chain in New Zealand to see there are a limited number of cold and flu medication brands on the market – as a result, even if supermarkets were competitive, there may be market power in the provision of these products to supermarkets which would lead to changes in the final price to consumers.
Green and Porter 1984 indicates what we may intuitively expect when an oligopoly exists in a market.
If the firms in the industry follow their own incentives to maximise profit they will compete and drive down the price – leading to all of the firms achieving lower profits. This is a traditional prisoner’s dilemma.
As a result, if the firms could collude to keep prices higher and act “like a monopoly” they could all be better off – since a monopoly sets prices in the market at the point that maximises industry profit.
The tension here is that each individual firm has an incentive to undercut their competitor a little bit, as the lower price allows them to steal some customers from their competitor and thereby increase their own profits – albeit at the cost of the profits of their competitor.
The question here is how might collusion breakdown if demand conditions change?
Green and Porter 1984 take a situation where the firms are colluding in a high demand state, and then the economy switches to a low demand state – for example the movement from winter to summer.
If an individual firm sees demand for their product drop there could be two drivers: Firstly, general demand for the product may have declined. Secondly, their competitor may have betrayed them and cut prices – thereby stealing some of the individual firm’s demand!
If the firm is uncertain about what the cause of the drop in demand is, then they believe that there is some chance they have been betrayed. In order to signal that they will “punish firms that renege on the collusive agreement”, our firm will have the incentive to cut prices.
The willingness to increasingly compete in low demand states is needed to ensure that collusion is maintained in the high demand states – by showing to everyone else that the business is serious about punishing firms that defect.
So our intuition and these models all indicate that cold and flu medication prices should fall in summer and rise in winter – which is the opposite of the result we find. So, Gulnara help me out here – why could this be?
This brings us to our sixth argument – tacit collusion and the temptation to cut prices.
Winter and summer are known quantities – we have them in our calendar, and although we don’t know the weather perfectly we have a good idea of what we will be doing during both.
In this way, a large portion of the difference in demand between seasons will be known. Furthermore, the actual price set by the other firm is observed – it is on the same supermarket shelf, on the website, and in the flyer! So it would be perfectly observable if competitors were cheating.
When the oligopolists have knowledge of what the other is up to, we get quite a different problem. Something that Rotemberg and Saloner 1986 discuss.
In this model a high demand state increases the “temptation” to betray your competitors to try to steal the entire market. Because of this, it becomes difficult to sustain collusion and collusive agreements will break down.
In our example this means that, during winter, there is a surge in demand for purchasing cold and flu medication. If our medication providers were able to maintain collusion, they would experience a big increase in profits, but the expected benefit from cutting prices (whether your competitor does or not) is now much higher. As a result, they cut prices and we get some sweet discounts.
Now you might say this sounds dumb – just don’t cut prices! But the key thing here is that all of the firms are responding to the change in demand conditions – not each other’s actions. If one firm does not cut prices, it is attractive for the other firm to cut prices. If that firm did cut prices, it is still the best response of other firms to cut their prices. Cutting prices is a dominant strategy!
In the oligopolistic setting all firms have this incentive when colluding – however, it is how the “benefit to defecting” compares to the “benefit of maintaining collusion” when a high demand state occurs that drives the result. Essentially, the arrival of the high demand state has increased the benefit to defecting while keeping the benefit of maintaining collusion (which are expected future profits) unchanged.
Although this sounds like a good explanation it has one issue – the high demand state is supposed to be unanticipated (as demand is modelled as a stochastic process). However, winter and summer are known!
Why does this matter? If the timing of the high demand state was known, and the oligopolists knew defection would occur in the high demand state, then there would be an incentive to deviate prior to this – changing the timing related to prices, and potentially unravelling collusion overall.
To rescue this model we would need to ask if there is some uncertainty about demand – and there is. The cold and flu season starts, ends, and peaks at different dates each year: Flu Season | CDC – here medication sellers will know when the high demand state has started, but there is some uncertainty about when it will start before the fact.
We’ve only noticed the strongly discounted prices when I’ve had a cold, which has been once this season has kicked off – as a result, it could be this form of “high competition during high demand states” that is driving the surprising result of lower prices for cold and flu medication in winter.
Although ECON101 supply and demand are powerful tools – they don’t tell us everything. The counterintuitive behaviour of prices tells us that sometimes we need to dig a bit deeper to understand what is going on – in terms of the cost of production, market structure, and structure of consumer demand.
Contrary to a lot of what you read on the internet, supply and demand is useful and powerful – but we just need to be honest that circumstances can be complex. As a result, keeping an open mind and always disciplining ourselves against the observed data is necessary for doing good economics.
If you are interested in thinking about this a bit more I did a quick google search after writing this script and found some nifty resources – they will be attached below. I’ll also add a few blog posts Gulnara and myself have worked on that chat about similar circumstances.