A couple of weeks ago, non-resident Samizdatista Alice Bachini pointed to this Telegraph piece in praise of supermarkets in general and the Tesco chain in particular, which explains that supermarkets help us to save time and money, make life easier (particularly for women), and provide a tremendous range of stuff much easier to buy at all sorts of odd hours if necessary, that Tesco provide a fine online service for people who want it, and that all round these are really good things and should be applauded. (Oddly enough, my fellow Samizdatista Jonathan Pearce wrote a similar piece between my starting and finishing this piece). Now this is a good article – it even takes a brief time out to denounce the Common Agricultural policy as evil – and on the whole I couldn’t agree more. However, there is one important issue that the author (Alice Thomson) missed. Midway through the article, she says the following
Supermarkets are always accused of sacrificing quality for quantity. Actually, because they buy in bulk and have a rapid turnover, it often means that their fruit is fresher than their competitors’. At Tesco yesterday, I counted six varieties of autumn apples, four from Britain.
Because they have such huge buying power, they can also take a gamble on exotic produce. The old corner shops were great for a packet of cigarettes, but they’d never have sold fresh basil. Supermarkets have made us more rather than less adventurous.
The result is entirely true. Supermarkets today contain a great many more lines than was the case a couple of decades ago, particularly fresh foods. But she is wrong about the reason. The reason why supermarkets are able to provide so much better products is not about buying power. It is instead almost entirely about the benefits that have been obtained by supermarket chains developing complicated computer systems to handle their logistics.
I will return to this, but for now a digression into economics. Economists in recent decades have often talked about something called the productivity paradox. Essentially, this notes that in recent decades businesses have been spending more and more money on computer systems, and statistics that measure national and personal wealth have not shown any improvement due to this equipment. There have been various speculations as to why this is so, but as the computer revolution has become more and more pervasive in all aspects of life, the idea that we do not gain any economic benefits due to computers becomes steadily more ridiculous.
To be fair, some benefits have started to show up even in the statistics in the last five years or so, but still these seem way too small, given that the world actually has been revolutionised. It seems that the numbers are wrong, but the question is how.
So we get to how the statistics are actually calculated. Basically, the statistics show an increase in prosperity if one of two things happen: incomes go up or prices go down.
When they are measuing just how much better off people are this year compared to how they were last year, economists generally look at GDP per capita. Essentially they look at the average income per person this year, see how it has grown compared to last year, subtract the inflation rate, and what is left is real growth in GDP per capita. For real GDP per capita to grow, incomes have to increase faster than the inflation rate, and for this to happen, workers have to become more productive – that is, they have to produce more goods and services per hour. Therefore, figuring out how best to encourage productivity growth is in essence the key task of economic management. (By the way, one great cause for uncertainty in economic statistics comes from the measurement of the inflation rate. Do this incorrectly, and stated numbers for GDP growth can be and often are wrong. Most economists believe that this has been happening for the last couple of decades, which is one reason why claims that “real wages have stagnated for blue collar workers since 1970” and the like are largely bunkum. But I digress).
The key point is that where new technology causes prices to drop, the statistics are good at showing it. However, where new technology changes the nature of the product, but the price remains the same, the statistics have much more difficulty showing it. And yet, the benefits are still real.
And that is where we get back to supermarkets. What the computer revolution has done for supermarkets is to allow them to create much more complex supply and distribution chains. This means that they are able to use a far greater number of suppliers, and that their stores can carry a far greater number of different lines, and that transport from suppliers to warehouses to shops has become far more efficient. (This leads to the biggest change in supermarkets over the last 20 years. The number of different lines they carry has increased by a factor of ten or more). They can keep track of approximately how many of each item is on the shelves of the supermarket at any one time. It is not necessary for people to look at how long perishable foodstuffs have been on the shelves, because the computers can keep track of this, too. If there is sudden demand for any particular item, then the store can be restocked extremely quickly in this item. Thus the supermarket has a far better idea of what customers want to buy, and it can improve its offerings. It is possible to stock 12 different types of apricot because detailed facts are known about how fast each type moves off the shelves. Because they know this they can make sure that the produce all remains fresh. Tesco’s computer systems have allowed them to increase the complexity of their stores dramatically, to the benefits of customers.
However, what they have not done very much is drop their prices. They offer a product that is much more tailored to their customers wants, and they provide customers with products that they may never have been able to buy in supermarkets (or anywhere) before, and they provide an all in one shopping experience, and customers are willing to pay for these benefits by paying higher prices than they might be able to get elsewhere.
Now, when quantifying the economic benefits of this, traditional ways of measuring look for lower prices. Tesco don’t provide lower prices, they provide a more complex product. Because people are paying about the same amount (or more) on groceries, and because Tesco are using about the same number of staff to provide it, the economic statistics do not demonstrate that anyone is better off, when obviously they are. People are buying things of the same value that they did before, but are buying much more complicated combinations of things. The economic statistics do not pick this up. But if you can see the value, one aspect of the productivity paradox is explained.
(This is one instance of a general effect of the computer revolution: the presence of computers allows companies to runs systems that are more complex in lots of ways. Another place this is evident is in the realm of design. Products design contains a low more style, a lot more choice, and a lot more complexity than was the case in the past, again largely due to the presence of computers. Virginia Postrel noted this fact a couple of years ago, and found it so fascinating she ended up writing an entire book about it).
Another different model for running supermarkets is the low cost supermarket model, in which a company offers a small number of lines as before, and no fancy service, but simply uses the improved logistics due to the computer revolution to track down and purchase from really low cost suppliers. This is the model of the two German chains Aldi and Lidl. In this case, prices drop and this does show up in productivity statistics. This is also a successful model, but most people seem willing to pay more for the more complicated service of the likes of Tesco.
In recent years, the retailers with the best computer systems have been able to expand their businesses in a way that other retailers can’t, and have had their stockmarket value expand dramatically. Once you have a really complicated computer system keeping track of all aspects of your business, you can do a huge amount with it. For instance, Tesco’s highly successful internet grocery business was relatively easy for them to set up, as it just piggybacked on the existing system. Britain has peculiar Sunday trading laws that only allows stores to open for more than six hours if they have a floor area of less than 280 square metres. Tesco have been able to fit full service supermarkets on that floor area, but could not do so without their highly complex inventory and logistics systems. Supermarket chains with good computer systems have been able to buy stores from chains with bad computer systems and have been able to change unprofitable stores into profitable ones just by changing the logistics and supply chains. (The takeover in Australia of many of Franklins stores by Woolworths is a classic example).
In fact, Tesco are so good at this that it is almost frightening. The British government is threatening to introduce compulsory ID cards, which will threaten our privacy and civil liberties in a wide assortment of ways. However, this is only the government. The government lacks the competence of a really good private sector concern with profits on the line. If Tesco were running the national ID card scheme, that would really be scary.
Except, of course, Tesco do run such a scheme. They call it a loyalty scheme, and give me a card to track my purchases, create a large customer database of my preferences and supposedly make special offers specially catered to me. I don’t have one of these because I value my privacy. (On the other hand, Tesco could still no doubt track me if they really wanted to, because I usually pay with the same credit card). However, this does illustrate the central dilemma of this type of technology. The benefits are potentially great, but the privacy implications are potentially great also. (We are also starting to see trials of systems in which every item in a store carries an individual radio identification tag. Once again this will help make supermarkets more efficient, and will no doubt help them minimise theft, but once again the privacy implications are troubling).
And computer systems aren’t everything. Britain’s number 2 supermarket chain is Sainsbury’s. And I have to confess I like Sainsbury’s more than Tesco. This is because they do really well for nice gourmet stuff (great cheeses for instance) and their wine and beer sections are full of really interesting stuff. Tesco are constantly praised in the business press and Sainsbury’s constantly criticised for not getting their logistics right, but they are still actually not my favourite place to shop. In my mind, Sainsbury’s do slightly better with respect to human things like differentiating between a good wine and a really good wine. Tesco’s computer systems are wonderful with the large scale stuff, but sometimes fall down slightly on the very small scale stuff.
This is all relative though. Sainsbury’s actually do all the computer stuff very well too. It’s just that Tesco do it better. Or perhaps I am just a snob.