Technology is going to change the way we work. In the research group that I work in, we have two Baxter robots. These humanoid robots are designed to effectively replace human workers in low-skill manual tasks, particularly in industrial environments. They’ve already been put in some factories in the US, particularly in small businesses. Earlier this year, researchers at Columbia University showed how a Baxter robot combined with machine learning could iron clothes. I published a paper earlier this year where Baxter could play chess against a human. Every day more applications are being revealed. These robots cost about US$25,000, which sounds like a lot until you consider that it’s roughly one year’s salary for a low-skilled worker. The robot doesn’t need to rest, it doesn’t get sick, it doesn’t need holiday pay, and generally achieves a lower error rate than a bored human doing repetitive tasks all day.
Automation, industrial or otherwise, is treated like the bane of the low-skilled worker. For researchers, simple tasks are the easiest place to start, and the technology already exists to make hundreds of types of jobs redundant – the main barrier is that it’s currently economically unviable to implement. The fear is that introducing this technology will cause mass unemployment and widening inequality. The traditional economic response to this is usually “well people can just upskill”, which is fine as a principle but not practical for many people. Upskilling requires education, education costs time and money, and not everyone has access to the necessary resources or support to “just upskill”. Even when everyone does upskill, there may not be enough jobs for all the upskilled people – just ask the legal industry.
The Universal Basic Income (UBI) seems like an attractive option to help alleviate the pressures on all people as their jobs become more uncertain and insecure. If we can provide every person in the country with a base level of income from the government sufficient for some minimum standard of living, then work is no longer a necessity for survival, but something that we do because we want to. We become more incentivised to find fulfilling work, not just work with sufficient income. Individuals become willing to take larger risks such as starting up small businesses or moving into unpaid education or volunteer work because the UBI safety net can catch them if they fail. Social welfare becomes a lot simpler, and the government doesn’t need to somewhat arbitrarily decide who is deserving of a benefit and who isn’t. I don’t necessarily agree that the UBI is the correct or only answer, but I accept that the UBI is one potential solution to address technological unemployment. However, if we are going to debate the merits of having a UBI we need to consider an important point.
My main concern with a UBI as a solution to technological unemployment is the nature of time. People losing their jobs due to advancements in technology happens gradually and in small pockets of society. In the past, this has been slow enough that humans have generally been able to adapt. When I say slow, I mean over the course of years or decades. Artisan weavers were replaced by mechanised looms, and while some people were hurt by losing their jobs, there was no broader societal upset. More recently, retail cashiers have been replaced by self-service checkouts, but this isn’t necessarily seen as a direct cause of rising unemployment. Most of these changes have been small in scale because technology doesn’t change jobs on a large scale overnight – the human element ensures that technology is introduced slowly and cautiously. Business owners are conservative, and don’t like risking the future of their businesses on some technological fad that might be outdated within six months. Technological unemployment is a continuous process, not discrete.
In most proposals for UBIs, the policy change would have to be instantaneous. One day, everyone is paying a certain level of tax and getting nothing (or a benefit), the next day new legislation kicks in and suddenly everyone is paying a different level of tax and getting a $11,000 a year. This implies that we have to identify some threshold where we say “okay, enough people have lost their jobs to technology, the UBI is needed now.” In the meantime, all the people below that threshold who have already lost their jobs will suffer. Perhaps we’ve already started to see this, with rising homelessness, rising inequality, and rising job volatility. When we apply utilitarian macroeconomics and tax policies, individuals fall through the cracks all too easily. We either have to wait for the problem to worsen and for some tipping point to happen, or bring in the policy too early and encounter unnecessary costs.
So perhaps, if we are going to have a UBI, what we need is unfortunately complicated – a gradual, slow increase of the UBI to match the gradual, slow changes to the labour market. Any policy implementation has to be discrete, putting it at odds with the continuous nature of technological unemployment, but at least making many small steps might be better than making one big step. So rather than jumping straight to $200 a week per person (and bankrupting the government), maybe we need to start at $200 a month or every two months. It’s not enough to live on, but it’s enough to start sending signals to society that the way we think about work and labour is going to change. Make it opt-in too so that people have to actively participate and understand what’s happening in order to benefit. We’d still have to initially keep some forms of social welfare benefits like jobseeker support (unemployment benefit) to help people transition between jobs (perhaps the amounts paid out for benefits can decrease as the UBI level increases). But over time, as technology makes our society more wealthy and prosperous, the UBI can increase sustainably to a level that supports all people in our country. It may feel like small drops in a bucket, but maybe that’s exactly what we need rather than pouring a jug of water into the bucket and watching it overflow.
Some would argue that this is difficult to do because it increases compliance costs, it becomes harder to educate people how this scheme works, and it become susceptible to over-reactions to short-term fluctuations rather than long-term trends. For some reason, when it comes to tax policy or social welfare policy we wait for ages and ages for changes to happen because we don’t want to confuse people by changing it too often. What I’m advocating for is a number of small changes more frequently, rather than one big change and then waiting a long time to revisit it. We already do this with the Official Cash Rate – we recognise that it’s important that for the Reserve Bank to react to the changing macroeconomic environment eight times a year rather than just once a year. Why shouldn’t we also be reacting to the changing labour market in the same way?
Of course, any movement towards a UBI requires a substantial change in how the government raises revenue too, whether that includes a true capital-gains tax on all assets, changing the tax brackets to increase the contributions of the super-rich, and/or creating incentives for multinationals to keep their money in New Zealand and pay tax rather than taking the profits overseas. The government’s expenditures would have to change too, with a hard look at superannuation, tariffs and subsidies, and social welfare more broadly too. But each of these policy changes should stand on its own merits independently, as well as together when considered in context. We can talk about whether a UBI makes sense and how to best implement it first, so that it can become a key plank of a model for how government can support society in uncertain times.