The relationship between innovation and inequality is rapidly moving up policy agendas. For years an implicit ‘trickle down’ theory dominated how policy makers thought. It argued that resources should be concentrated where they could achieve most – on the top universities, the highest growth firms and the ones on the forefront of technology. If high-tech, highly innovative firms and sectors grew, that was bound to result in more wealth and prosperity for everyone else as opportunities trickled down from the dynamic centres. That argument helped justify more funding for R&D, subsidies for venture capital, and a clutch of policies to promote clusters of all kinds.

I’ve been quite closely involved in these ideas over several decades. So it’s difficult for me to admit what is becoming increasingly obvious – that parts of the theory were substantially wrong.

"For customers at Walmart or Amazon there have been very tangible benefits. But for workers, the results have been disappointing to say the least."

If you’re in any doubt, just look at a few of the poster children of innovation in the modern world. Silicon Valley is the standout example, as successful as ever at generating vast wealth, and a continued stream of new technologies. All of that is great if you’re a Mark Zuckerberg, or a venture capitalist. But it’s pretty hard to show much evidence for the trickle down.

In the US median wages and salaries have been stagnant or even falling over the last few decades during precisely the period when clusters like Silicon Valley have boomed. For customers at Walmart or Amazon there have been very tangible benefits. But for workers, the results have been disappointing to say the least.

Israel is an even more intriguing example. It has done extraordinarily well in recent decades in growing its export oriented IT industry, and other countries have spent the last decade wanting to copy the successes of programmes like Yozma. What’s much less well known is that there has been no measurable benefit for most other sectors in Israel, and no evidence of much, if any, benefit for the majority of Israeli citizens. As in the US, the growth areas sometimes look more like islands than like engines driving the rest of the economy.

In London, the city where I work, the pattern is also clear: booming high technology sectors, in digital, life science and the creative industries, but relatively little evidence of much benefit for a high proportion of Londoners who have seen their opportunities, and income stagnate as house prices have soared.

Across the western world, city ministers and mayors regularly find themselves having the same conversation. Yes, we now know much more about how to cultivate buzzing creative industries, universities, knowledge intensive industries and so on. But we have almost nothing to say to around half of our population who face the prospect of bad jobs or no jobs, and look on with dismay and envy at the windfall gains accruing to the elite insiders.

The links between inequality and innovation aren’t straightforward. The evidence just doesn’t support the two most common claims: on the one hand the claim that inequality is necessary for societies to innovate; uneven rewards motivate people to take risks, create and invent, and on the other hand the claim that innovation depends on tapping the talents of everyone.

So what is to be done? I’d suggest four main avenues to explore.

First, it’s clear that some industrial systems do much better at spreading the opportunities than others. Germany remains an engineering and technology powerhouse but has sustained a more equal distribution of pay and chances. There are many factors in play, from apprenticeships and training at work, to corporate norms. The rest of the world can’t simply copy Germany (though both the UK and US have recently copied their Fraunhofer Centres) but there are important elements of the German system which should serve as prompts.

"It’s clear that some industrial systems do much better at spreading the opportunities than others."

Second, discretionary funding for R&D of all kinds is bound to follow the interests of those with a say. In the US venture capital has a far stronger voice than trade unions. In some countries (including Germany, again) the opposite is the case. So if you want innovation to serve equality then you need a more open and inclusive involvement in deciding how scarce public money should be used. Concorde is the prime example of what not to do: vastly expensive innovation programmes designed to serve the needs of a tiny minority of the very rich, rather than the needs of the majority for better energy, public health or transport.

Third, the habits and mindsets of innovation need to be widely cultivated. If children and teenagers get experience of practical problem solving, entrepreneurship and technology from a younger age they may be more likely to thrive (one reason why I’m so keen on Studio schools).

Finally, there needs to be a more inclusive dialogue about what innovation policy is for. Our research last year showed that the UK population was strongly supportive of new ideas, but that the standard rhetoric of technocratic innovation policy – and of ministerial speeches all over the world – simply didn’t chime well with their values. For the public it matters much more what innovation is for, and questions of ethics and of who benefits count in ways that seem to be ignored by many innovation agencies.

These issues are set to become ever more central to public policy. The clashes in San Francisco over private buses taking the employees of Google and other firms to work are a symptom of the wider problem. The rest of the public have started to question the trickle down thesis which has dominated debate for a generation now. We need more research to understand what’s happening and why some countries do better. But we also need a new model of innovation policy that is designed to be inclusive both in widening participation, and in shaping whose needs are met.