Dan wrote me a comment on my post on modelling local economies and the effect of shops which generate more income but send profits outside the local economy. It’s quite long so I’ve put most of it below the fold. Some context may be found from this post i’ve linked to before, by Dan at Indymedia.org.uk, about the
redevelopment plans current for Burngreave, Sheffield. Even if you’re not interested in redevelopment policy, there’s stuff about the utility and use of simulations that has general interest
Some abbreviations i’m not sure he defines: LM3 = Local Multiplier 3, a measure developed by the NEF which gauges how much of money spent in the local economy stays in the local economy. NEF = The New Economics Foundation. ABM = Agent Based Modelling. ODPM = Office of Deputy Prime Minister.
Anyway, Dan says:
This is all a bit like wading through underbrush at the moment. One day in the future, the concepts we’re trying to get at may emerge from the murk, but for now….
1. The value of modeling
2. A human network syndrome?
3. Capitalism, network breakdown
The value of modeling
Take a look at http://www.econ.iastate.edu/tesfatsi/abmread.htm for a nice summation of some of the epistemological wotsits around modeling.
One goal of agent-based modeling, the site says, is normative understanding – ?How can agent-based models be used as laboratories for the discovery of good designs? ABM researchers pursuing this objective are interested in evaluating whether designs proposed for social policies, institutions, or processes will result in socially desirable system performance over time.?
Well, that’s tricky. What counts as good? And, from that, what counts as a good policy? A huge study was recently concluded in Chicago ? absolutely excellent conclusions, but one of the policy recommendations was that public housing be destroyed and replaced with more home-ownership. It was understood how important a factor housing is, but no-one asked ‘what is it about the housing that’s causing the problems?’
Milton Friedman came up with a rather curious little theory. All hypotheses have assumptions, he says. But he claims the conformity of these assumptions to ‘reality’ is not a test of the validity of the hypothesis: the test is its predictive power only. So, as an example, any theory that has a ‘rational agent’ as its foundation cannot be falsified on the grounds that people aren’t like that. As Hollis puts it, “all models involve abstraction, and the only test of the merit of an abstraction is comparison of the resulting predictions with experience. As it turns out, the predictions implied by perfect competition models fare better.” Friedman suggests that all assumptions within theories can be held to be ‘as if’ true. This needs reiterating: it is their predictive power, not their veracity, that should be the criteria for their acceptance or rejection. Hollis argues that “the significance of this ‘as if’ is that it lets Positive science dabble in unobservables, provided they are not thought more than useful fictions.”
From this point of view, ABMs are all just ‘useful tautologies’. As is economic theory that sees trade in financial services in the same light as trade in wage labour.
But with this little LM model, its not predictive power that I’d say was important. Even if it had that power, it would be fairly trivial. What it’s done is much more exciting: rather than having to bribe a cohort of people to play the agent roles, Tom’s allowed a very quick ‘what if’. It’s a bold conjecture ? individual betterment can be locally impoverishing. That begs the question: given that an area could itself be an agent, might the effect cascade? Where is the balance of stability in the system? How does that balance relate to the relative weight of micro-and-macro wealth?
If modeled, all these questions would still be useful tautologies only. But would we want to make a claim for their predictive power?
Bold conjectures can sometimes falsify. If we go all Lakatos briefly, maybe the right conjecture can act as a spanner in the protective belt of standard economic theory ? one that, if wiggled and wiggled, might gain us access to the hard core. A bit like the Matrix. Um. Grandiose messianic fantasy on my part, there ? sorry. Back to theory.
The spanner-in-the-works potential of LM3 has always drawn me to it. Using ABM to dig into its implications further could be very cool. But back to this ‘useful tautology’ ? prediction / analytic-synthetic thing…
Economists have a knack of concluding, when a theory looks a bit poorly in the light of prediction, that a country had not xed enough, or too much, or xed too early, or too late. Replace x with whatever international financial institution policy is in vogue this week. (Very often used by the World Bank to shore up its theories. The Bank?s reasons for failure of its reforms tend to be that they were either implemented too quickly, too slowly, or imposed by government elites and foreign donors, causing a political backlash, or they were reforms hijacked by special interests. Certainly, the problem is never the reforms.)
This will always, always be possible. Any claims for predictive power ? comparing economic theory to something like the theory of gravity – are simply rubbish. There is way, way too much scope for ad hoc-i-ness in there.
So what might this mean for our little LM venture?
One ? ABM is powerful because it can mirror the world. The best example of that is a recent find: ‘the game’. (See http://www.icosystem.com/game.htm) It works in Java, and it does exactly the same when you try it with a bunch of people: a nifty little real-world example of an emergent property. And, in the real world, the fact that it’s human beings ? hugely complex creatures, after all ? enacting the rules makes no odds. The dynamic works. Trivial ? but a point worth making, because while a lot of us digital natives clearly accept the parallel (and the principle of ‘substrate neutrality’), to others its simply an horrific idea, and one that gets people all vexed about free will.
Let’s have another go at articulating that point. There’s an old division between rationalism/deduction and empiricism/induction. The former is based on looking inward; on building logical conclusions. The latter on finding regularities in the world ‘out there’. On the face of it, it looks like ABM ? or perhaps more precisely, the raw number-crunching power of computers ? makes a mess of this intuitively useful distinction.
Why? Modeling means rationalism: your system is as logically tight as any syllogism. It has to be. But then, it means empiricism ? let it run, wait and see…
The LM model isolates one possible dynamic, as does Schelling’s classic segregation model. But is that all it does? Do we only get ‘well, it could be like this…’?
Christ knows. I’ve mentioned ‘the game’. The rules there work in both the virtual and the real-world case. In the real world, all other factors to do with how complex human interactions can be just don’t count. But how do we know when they might, and when they might not count?
With the LM model, a simple proposition than thus lead to big questions.
Two – and much more importantly – once this parallel is conceded, we can then make links with the empirical much more powerfully than Friedman-esque tautologies. If the parallel is conceded, then there is less need for anal retentiveness about the analytic-synthetic distinction: ABM becomes a reflexive tool for poking around in reality; reality helps feed back ideas and offers empirical hypotheses. Like –
A human network syndrome…
Here’s a recent story in the Guardian – http://www.guardian.co.uk/comment/story/0,3604,1514200,00.html –
Its talks about the ‘adult surrender of public space to anyone willing to scare others’. In it, Jenni Russell laments that, when someone is harassed or attacked in a public space, they can expect no help. Everyone around is ?scared, or indifferent, or embarrassed, or afraid of being accused of assault themselves. Nor does anyone have any confidence that, if they do intervene, they will be backed up by others.?
She goes on to argue that:
?the police cannot prevent crime, intimidation or harassment. All they can hope to do is occasionally catch a perpetrator. They are meant to be the enforcers of the rules that we as a society have decided to live by – not the only upholders. If we see people being mistreated, then surely we cannot just turn away and thank God that we are not the targets. If that is what we are doing, then we are starting, literally, to become less civilised. From Hobbes onwards, we have understood that civilisation means the replacement of the power of the aggressive individual by the power of the wider group.?
Quality is much more difficult to use than quantity ? especially if you’re wanting to convince policy-makers of anything. The Guardian article above has the ‘intuitive’ factor. It may feel right ? but possibly only because of nostalgia.
It must be time to bring Putnam in. Bowling Alone is, in my opinion, still one of the most radical books ever in the history of people writing things down. Why? Well, first because we always make grand claims for the books we’ve read because it makes it sound like we’ve read more than we have. But also because it gives such a quantitative grounding to social networks: not in an abstract way, but connecting social capital with social goodies. It gives us a heap of powerful tools for measuring the good things that social networks produce. And it makes empirical statements about human network nature ? ones that you wouldn’t just hypothesise out of thin air.
Putnam uses many, many different indicators to build up his picture of declining social capital. LM3 itself is also an indicator. What we may be able to do is come up with a list of ‘social goods’ and think about how LM3 might correlate to these goods, and what role ABM might play in investigating the connections and looking for new dynamics.
The core idea of social capital theory is that social networks have value. This value is a function of one’s connectedness – ‘social networks and the norms of reciprocity and trustworthiness that arise from them’.
Now, an important empirical hypothesis relates to reciprocity. ‘Generalised’ (as opposed to specific) reciprocity means ‘I’ll do this for you in the confident expectation that someone else will do something for me down the road.? Putnam argues that “a society characterised by generalised reciprocity is more efficient than a distrustful society, for the same reason that money is more efficient than barter. If we don’t have to balance every exchange instantly, we can get a lot more accomplished.”
His hypothesis is thus that “frequent interaction among a diverse set of people tends to produce a norm of generalised reciprocity.”
This is unfortunate sociological garble ? but it’s intriguing and important. What Putnam’s done is propose that, if you increase human interaction, you’ll produce a community-wide emergent property ? one that means people within it will do things for each other because they are members of a community.
He does break down those interactions, but for our purposes its enough to note that they must be substantive, and they usually happen in established forums ? whether church, work, local stores, house visits, etc. Paying for something in a supermarket doesn’t build social capital (much!)
Putnam also heaps on the normative: why is social capital so important? It correlates to health (mental and physical ? people with more social capital live longer, for example), wealth, community safety and local democratic / economic vibrancy. I highly recommend reading the ‘so what?’ section in Bowling Alone. Many, many pages of tightly argued, thoroughly backed up reasons why social networks equals good things.
(One little example: social capital has a disproportionately positive effect on the job prospects of people in less well-off areas. One reason is the phenomenon of ‘weak ties’ ? you are more likely to find work through an occasional acquantance than a close friend. Access to networks thus means access to the information and back-up you need to take part in the local economy. Again, it should be clear what difference the LM3 score could make here ? and it should be easy enough to check for a correlation.)
Putnam says that ?the causal arrows among civic engagement, reciprocity, honesty and social trust are as tangled as well-tossed spaghetti? however, we need to recognise that they form a coherent syndrome.”
If this is true, its a syndrome that increases with the level of local interaction. But we’re talking about LM3, aren’t we? Well, I’d like to hypothesise that the LM3 score positively correlates to social capital – and therefore to all the other goods that social capital brings.
And, somewhere in the back of my brain, there’s a little goblin going ‘mmm mmm! Nnff! Mm-nn!’ I can’t quite hear what it’s saying, but this idea of a syndrome, and human network nature ? and the fact that Tom’s got hold of one little quantitative property. Individual betterment can be locally impoverishing… if an area is itself an agent, might the effect cascade? Where is the balance of stability in the system? How does that balance relate to the relative weight of micro-and-macro wealth?
We’ve got a series of goods ? economic well-being, health, democratic vibrancy ? that share common features, and are mutually supporting.
And vitally, we’re not stuck in having to talk about ‘useful tautologies’. For example, we can ask: does the food economy have a different impact from primary products? Is it more important to the health of ‘the network syndrome?’ How effective an indicator is the LM3 score? (It may be that LM3 for food works well, but not for steel or call-centre labour.)
So what have we got here? A way to look at the economic and the social in a new way, that allows much more nuance? Well ? I imagine it’s been done, but I wouldn’t want to presume that it has. And I certainly know that, when it comes to regeneration, the powers that be are still letting Tesco et al do their research for them, defining poorer areas as ‘under-served markets’ and proposing, er… Tesco and MacDonalds as the solution! This alternative is worth pursuing…
LM3 as an indicator
Back briefly to some questions about LM3. If the money’s moving between local businesses and contractors, all very well for them. How does it actually benefit, say, someone who can’t afford to buy decent food? The local multiplier implies that the local economy is richer, but how widely is that benefit spread? I’ve made a grand claim that it might correlate to stuff, but it might not.
First: if its an indicator, then it’s one way of measuring how well an area’s doing. But if we want it to do any more work for us, it gets trickier.
Lets imagine that we could change tax incentives so that local people and businesses were encouraged to trade locally ? for example, by a VAT exemption on locally produced goods. Does this mean we could socially engineer an increase in local capital circulation ? and thereby increase its correlates? Not necessarily. Whilst LM3 might be an indicator, it may not be a very effective cause. This goes back to the discussion earlier: what actual value is it to a community if the take-away is trading with the landlord, who’s trading with the hairdresser? A high LM3 score might be symptomatic, but it won’t necessarily be any guide to action.
Or maybe it will. It seems obvious that trading in local stores would correlate with a higher levels of interaction, and we already assume that interactions equals social capital equals local social goods. But there’s work to be done, ain’t there?
Also, both Putnam and the NEF suspect that local networks can reach critical points where they will break down much more quickly ? phase changes will take place. Putnam sees this happen in networks of activism, where the level of involvement drops below a certain level and starts to positively feed back. NEF sees the same happen with local shops: a tipping point is reached where the fabric of local shopping cannot hold together. We don’t know at present what factors are most vital; which ones might act as bulwarks against them breaking down. But it may be suspected that the LM number might indicate it: again, it’s measurable, so we could find out.
Bringing LM stuff and social capital together also suggests another point: could generalised reciprocity be applied in the economic sphere? It certainly can for food ? people share food more when there’s more social capital. They are also more able to exchange, borrow, do odd jobs ? as well as take part in ‘black market’ activity, which even the ODPM has noted can have a positive role to play.
So, as an example, we have a way to look at whether local *food* networks specifically are more important than others.
Incidentally, this would be a powerful argument against those who accuse us of being recividists who want to return to protectionism. Response: yes ? let’s keep our trade in primary products across continents. It’s a really effective war deterrent, as the EU demonstrates. But where globalisation of certain sectors ? like food ? is damaging to the human network syndrome, then let’s keep it local. Best of both worlds. Oh ? and only use solar-powered transport!
It would be the parallel to Putnam’s “bonding social capital [which] constitutes a kind of sociological superglue, whereas bridging social capital provides a sociological WD-40.” That’s an interesting comparison to the multiplier effect. Why should local money networks ? ‘bonding money’, as it were – be better than ‘bridging money’? Maybe neither ? maybe it’s a balance.
So what of this syndrome? What of the balance between individual / micro and community / macro wealth? When I originally wrote about Bowling Alone, I said:
?Putnam has offered a way of defining and thinking about the individual in terms of the depth and extent of social connections. This is an important quantifiable way of breaking with the ?homo oeconomicus? model of the self. (Williams.) There are a set of ?goods? ? in terms of well-being, health, wealth and ability to have a democratic voice ? that can be directly correlated to the level of connections that self has. Notice that this still maintains the importance of the individual, at the same time that it emphasises the context of that individual. He understands these in rational terms as ‘self-interest rightly understood’ ? i.e. enlightened self-interest. The statistical basis of the concept of social capital gives it a great deal of weight, precisely because it gives the concept a grounding in a material measure. This also means that finance capital becomes intertwined with social capital.?
[I’ve already mentioned the Project on Human Development in Chicago Neigbourhoods [http://www.hms.harvard.edu/chase/projects/chicago/] Take a look for a different but similar approach to local networks, using the concept of ‘collective efficacy’.]
Capitalism, network break-down
We’ve been living in an era where even capitalists use the term capitalism. Everyone accepts the definition. But is it actually of any use?
I’ve argued that networks have value ? value that can be measured. It’s a hugely complex picture in some ways, but this basic truth is not difficult to grasp.
That’s what’s so great about Tom’s LM model ? it’s a very simple network model that highlights a simple truth. If you just slightly alter the way you think about wealth, sticking it in a network model, you find an important conclusion: individual pursuit of wealth may damage a community. Equally it may benefit it, but it’s the network nature of the wealth that matters, not the quantity alone. A policy that promotes local circulation of wealth may thus have a whole basketful of socio-economic benefits.
Putnam does not believe that the blame for decline in social capital can be laid at the door of capitalism: “the problem with this generic theory of social disconnectedness is that it explains too much: America has epitomised market capitalism for several centuries, during which our stocks of social capital and civic engagement have been through great swings. A constant can’t explain a variable.”
However, Putnam is more open-minded about the possible damaging effects of the globalisation of this same capitalism; he worries about the “replacement of local banks, shops, and other locally based firms by far-flung multinational empires”, but mainly because it means the disengagement of local philanthropic business elites.
Martin Wolf is a chap who used to be very left-wing, but is now a social democrat with a burning anger at anti-globalisation protestors. He sees globalisation as a positive, socially progressive force, and anti-globalisers as ill-informed wreckers. His book, ‘Why Globalisation Works’, is an attempt to argue with them. (And he’s a lot less patronising than Bhagwati..)
Early on, he uses the example of the US: no-one would claim that US states should close their borders and stop trading with each other. That would be madness. Therefore, in what way is it different for the rest of the world?
A Lot of the rest of the book tell us about the inescapability of markets; their status as natural human systems and the folly of central planning, etc.
But what if anti-globalisers aren’t against the connection of the globe?
From everything I’ve said here, what I’m personally against is the local dissolution of human-scale networks, which are being replaced by something much larger ? and, yes, more global. Markets, self-organisation and networks are part of what it is to be human. But they are not separate from all the other aspects of humanity I’ve talked about ? local wealth, health, community and self-determination.
In terms of policy – the normative task we started with – it’s problematic. A key property of networks is that they develop over time. Any radical intervention is liable to disrupt them. Is it possible, then, to think about how we might be ‘social gardeners’ rather than social engineers? Allowing the garden to grow by itself, but providing the framework…
In 1973, Schumacher was saying:
?In the current vocabulary of condemnation there are few words as final and conclusive as the word ‘uneconomic’. If an activity has been branded as uneconomic, its right to existence is not merely questioned but energetically denied. Anything that is found to be an impediment to economic growth is a shameful thing, and if people cling to it, they are thought of as either saboteurs or fools. Call a thing immoral or ugly, soul-destroying or a degradation of man, a peril to the peace of the world or to the well-being of future generations; as long as you have no shown it to be ‘uneconomic’, you have not really questioned its right to exist, grow and prosper.?
The strong connections that make up the syndrome I’ve been describing may help to kill the idea of ‘uneconomic’. It’s easy enough to have a sense that ‘everything’s connected’ ? and everything is, of course! But if it’s possible to demonstrate, first, the embedded nature of the economy, it’s network nature, and just how it connects to the human network syndrome of economic-democratic-physical well-being: we win!
ABM should be a powerful ally in this role. Tom’s model, for example, suggests it may be ‘uneconomic’ to allow personal enrichment at the expense of the community…
As a start, if we could investigate the relative importance of local food networks ? and the importance of ‘food sovereignty’ ? then we’d be taking an important step away from the two ideological poles of market fundamentalism and old-school millenarian socialism. And, I hope, a step away from the fascism that both of those ultimately embody.