Migration and wages: A model that is wrong but useful

The link between migration and wages is complex and confusing – especially when it is often communicated about in different ways (eg are we talking about wage growth now with regards to monetary policy, wages in specific industries due to the changing make-up of the economy, or long-run real wages?).  And I can’t be much help here.

However, I think this is one place where carefully using the macroeconomic model taught in ECON101 can help us to think about the issues a little bit – especially if we are narrowing the question to only “what is the monetary policy consequences of changes in migration flows“.  Now this model is wrong, assumptions in it are wrong, the outcomes it describes aren’t forecasts – but it clearly articulates tendencies we observe following a change in economic circumstances which will hold in more realistic models, and clarifies assumptions that may make these tendencies false.  We have pointed at this before for monetary policy – but lets outline a bit more now.

It is a model for thinking about the potential consequences of something in a critical way – not something that we accept uncritically as truth.  To me this is pretty damned useful as a way to start thinking about something, so let’s do it!

The basic ECON101 macroeconomic model takes our friendly circular flow, with households as the group that demands final goods and services and that supplies factors of production, and firms as the other side of these markets (a set of technology that creates final goods and services from factors of production to “maximise profit” in some way).  Sweet.

So when we talk about Aggregate Demand we are talking about all sources of expenditure on final goods and services produced in a country at a given price level.  Recognising that income=expenditure in our circular flow, and the AD curve tells us what level this is consistent at for a given price level, we are not talking about the idea that people just want stuff at a given price level (microeconomic demand) – we are simply saying that, if we have that price level and that level of output total expenditure of households is equal to the amount produced by firms.  Still, we won’t dwell on this here.

When we talk about Aggregate Supply we are thinking about the “supply” of factors of production.  When we put it this way the typical upward sloping Aggregate Supply curve seems strange – and truly it is!  So lets think about what we are talking about here.  With AD we stated that we were looking at the choice of households to demand and firms to supply output at a given price level, this is “equilibrium” for us.  But we don’t know what point on this curve gets chosen!  To get there we need to consider that there is also a choice that depends on the actual decisions of firms to buy factors of production and households to supply them.  As the price level (relative to factor costs) rises firms are willing to make more – cool.  But firms also need a way to source more factors of production at that fixed price.

In a sense, when we have an upward sloping AS curve we are saying that, as prices rise and real wages fall, the amount of labour purchased must rise.  This is why the concept is a bit unintuitive.  This is also why ideas like “nominal illusion” and “unsustainable changes in employment” are used – and why data such as overtime hours, delayed retirement, and the gap between expected wage inflation and inflation are used to judge what is going on.  (Note: Imperfection in the goods market gives us a great way to understand why a shift in AD may lead to firms increasing output for a fixed price – as it creates a “kink” in their MR curve.  But although this increases labour demand, we need to understand why more labour is getting hired – is there rationing in the labour market, is their some trickery, or is their rationing due to imperfect good market competition and price expectations – this is all I mean by AS being a bit of a mind freak 😉 ).

Ahh I said expectations – expectations make all this stuff more difficult, so we are leaving it to the side while noting that adding expectations is incredibly important for considering these issues.  In everything discussed below inflation expectations are fixed at zero, and no consideration of future expectations of output/income are taken into account.

However, this isn’t a lesson on the curves – Wikipedia will outline enough of that, or we can all just go along to an ECON101 course for a year.  Instead, given that outline we want to think about migration while accepting the concepts embedded in these curves will capture real tendencies in the economy:

Migration AD-AS

Migration is a positive shock for AD – the AD curve will shift right.  This implies that, for a given price level, the equilibrium level of output in goods and financial markets is higher.  Think of it this way, someone moves here and they now want to buy things irrespective of their income – they need to cut their hair, they need to eat.  This lift in expenditure leads to the process that shift AD right.

This shift then leads to more output for the same price level – except that firms will be deciding to change both prices and output due to this change.  As a result we need to consider the idea of what firms will produce given the factors of production on hand.

That’s cool, but migration increases supply as well right – these people work!  Yes, yes they do – but slow down for a second.  What is our AS curve telling us?  It is the price and output combinations of firms that maximise profit.  We can simplify this idea by just saying that firm set a fixed margin on top of the cost of the final unit they produce when setting price, that labour is the only factor of production that changes in the short term, that wages are fixed but since the marginal product of labour declines as we make more the marginal cost of production rises.

Got that, yup those are the assumptions that give us our AS curve.  If wages are fixed how do we get more people to work … well that is where we need to include short term assumptions.  We can get people to delay retirement, people working overtime, we can trick people who are looking for work into thinking this is a different job than it is.  None of these things are sustainable, but they can boost output now.  Another thing we can include is saying that we do have wage growth, but it is slower than price growth, and labourers don’t realise that – so they think they are being offered a higher real wage when they are being offered a lower one (nominal illusion)!

Now if the wage is currently fixed, what happens to AS when we have an increase in the population!  Well we need to split this curve in two a “short run” relationship where wages don’t change, or where wage growth doesn’t change at least, and a “long run” relationship where factor costs adjust until they have moved as much as prices (equilibrium with flexible prices and factors).  We’ll call these short run AS and long run AS – or SRAS and LRAS for short.  I realise that it is common to teach LRAS as just the top end of an overall AS curve but this is misleading for two reasons: 1) output can rise above potential, and LRAS is potential output, 2) shocks can influence LRAS and SRAS differently as patterns of production change following the supply shock, having two curves allows us to discuss that dynamic and the transition paths.

So we’ve said that AD curve will head right, but what happens to the SRAS. Well nothing at the old equilibrium point, that point just stays chill as firms are producing and using the factors of production they want to.  What about when prices rise, well those points don’t change either as the “wage rate” is fixed on this curve (capital is fixed, so DMPL exists irrespective of the increase in the stock of labour).  Instead what matters is how the process of wage changes comes into play! [Note:  The idea that it makes the SRAS curve flatter at some points as it changes their ability to actually hire staff at a given wage is a fair one – and really illustrates how ugly a “fixed wage” argument really is for a lot of these discussions 😛 ].

Given that we have a tendency for rising prices and increasing output.  This leads to a process that will eventually push nominal wages up, shifting our SRAS curve left!

But here is the kicker, the increase in the working age population due to net migration has also increase labour supply … moderating the increase in wages due to this process.  Overall the SRAS curve can go where-ever it likes as long as it chases down our LRAS which has shifted.  The question is then, which shift was bigger due to migration, the AD shift or the LRAS shift.  Output rises overall, who knows what prices do, and the change in output per person depends on the relative productivity of migrants and the response of the economy to higher scale.

There is more factors of production now, so more stuff gets made.  But the price level change is uncertain.

In NZ, economies of scale and our “high skilled” migration policies generally hint at migration increasing long-term incomes (the third bullet on this does some discussion there).  This model tells us nothing about this, even though it is arguably what many people talking about it care about the most.

But the louder debate has been about monetary policy and current wage rates. And to discuss that we need an idea of the relative size of these shocks and their timing – why have the “pricing pressures” we would expect from a sudden increase in the population not occurred, and how is that consistent with other economic indicators?

In other words, why hasn’t there been inflation from such a big migration boom in New Zealand!

Monetary policy

Monetary policy shifts the AD curve, so following a migration shock the central bank will be looking to get that curve to match these shocks without undue variation in output and prices.

But the model we just described points out a situation where migration will increase pricing pressures and lead to a process of wage growth kicking off until migrants are well integrated into the labour market.  However, from 2014 the RBNZ noted that migration was holding down wage growth.  By 2016 they gave more of a description of how this may be the case.

How do we get there?

  1. Employment based migration is boosting productivity/leading to rapid integration in the labour market (is this leading to higher MPL workers directly, or is there a corresponding capital flow – human or capital – which is used immediately?).
  2. Quick integration of people with lower wage expectations – if the increase in labour supply due to migration exceeded the lift in labour demand once wages start to adjust, the wage and price pressures may be downward rather than upward in the transition to the long-run.
  3. The return on New Zealand citizens who were already well integrated.
  4. Monopsony in labour markets:  If there are monoposony qualities to the labour market then firms WANT more workers at the current wage.  As a result, the increase in labour supply can immediately translate into higher output.

These may have been the factors in the structure of the economy that changed the transition (in terms of inflation and output) relative to what the RBNZ expected.

Another potential explanation comes through expectations – there was a recognition that individuals moving into the country would eventually boost supply, and so people did not expect inflation to rise, or may have been unwilling to shift prices in opposing directions over a year.  However, the reason I go straight to a labour market explanation is because of the high migrant employment – and high participation rate during all this.

In this way the RBNZ when describing this phenomena was NOT saying “migrants and holding down wages in New Zealand”. They were saying “the expected process of inflation following a migration shock did not occur as the structure of the economy differs from our expectations and experience, as a result inflation outcomes were lower than we expected and if we had known interest rates would have been set lower”.  That is a perfectly reasonable point.