Limits of supply-side reform

Philip Hammond says he plans to use Wednesday’s Budget to get us “match fit” for Brexit. This reminds me of an old error of Chancellors and political commentators – their habit of under-estimating just how damned difficult it is to increase long-run growth.

In fact, John Landon-Lane and Peter Robertson have shown (pdf) that the data are consistent with the possibility that national government policies have no discernible impact upon long-run growth at all. This is because most developed economies grow at much the same rate as each other, give or take a standard error and pathological exceptions such as perhaps Italy or Japan.

Dietz Vollrath has shown why this is. It’s because economies are very slow to converge to higher levels of potential output, even if such levels can be achieved.

To see why, imagine something happens to raise the level of potential GDP growth – say, some technical innovation or your pet policy gets implemented. For this to affect actual growth, companies must first see new profit opportunities, then persuade financiers to back them, and then retrain workers, install new equipment, hire more staff and so on. But there’s mountains of things that can go wrong here: new technology might not work well; it takes time for workers to learn new jobs; managers can lose control of costs as they focus on expansion; and so on. All these problems throw sand into the wheels and stop the economy adjusting swiftly to its higher potential level.

Resources – capital and labour – are not very fungible. One of the consistent errors of free market economists has been their unwillingness to see this.

But it’s not just them who make this mistake. In a new paper Charlie Cai and colleagues show that equity investors tend to pay too much for growth stocks because they under-estimate the difficulties firms have in expanding: expansion usually entails falls in asset turnover and in margins, facts which themselves limit that expansion.

This micro evidence fits the macro evidence that convergence towards higher potential output is slow. It also tells us that the inability to appreciate this doesn’t arise merely from motivated reasoning and political ideology, but rather from a widespread cognitive bias – a tendency to over-rate the importance of salient things such as policy announcements or current corporate growth and so under-rate the importance of the countless low-level barriers to growth.

For me, this has two policy implications.

One is that Chancellors should worry less about raising long-run growth and more about getting macroeconomic policy right. Failings in the latter can easily more than offset good supply-side policies: for example, Nicholas Oulton’s assessment of the Thatcher reforms was that “microeconomic success has been masked by macroeconomic failure.”  As James Tobin rightly (pdf) said, “it takes a heap of Harberger triangles to fill an Okun gap.”

Secondly, although the adjustment to higher potential output is slow, that to lower output might not be: companies can quickly close factories and shed workers. Economic policy should therefore obey a form of the Hippocratic oath: first, do no harm.

With this government committed to both fiscal austerity and a hard Brexit, it is disregarding both these principles. As our media is more likely to focus upon Hammond’s cheap talk than upon hard economic realities, it will probably not sufficiently be blamed for this.

Comments

[Patrick Kirk]

If John Landon-Lane and Peter Robertson are correct that national government policies have no discernible impact upon long-run growth at all, does that mean things like Brexit and TPP are irrelevant to the economy and jsut about identity signals?

[Luis Enrique]

On resources non-fungible point, you might like this book by Ricardo Caballero:

https://mitpress.mit.edu/books/specificity-and-macroeconomics-restructuring

"Caballero argues that macroeconomic models need to be made more "structural" in a precise sense and can not be maintained on the assumption that decisions are fully flexible. What is needed, he proposes, is the notion of specificity — the idea that factors of production are not freely interchangeable."

[David]

My feeble mind "Boggles" hopelessly trying to imagine a budget that is a "match-fit" for "Brexit".
A man with a mission indeed.
Maybe he should be carrying a jar rather than a box this time, and take the moniker.... Pandora.

[Ralph Musgrave]

So Chris, you're saying politicians are not to be trusted. I'm shocked...:-)

[Keith]

It has surely been obvious for decades that the only thing Chancellors should aim for is to avoid large falls in aggregate demand persisting for long periods. The experience since 1945 is that it is easy to reduce growth by mismanaging demand but there are no effective ways to increase economy wide productivity growth except of course for the ones that work all the time whichever party is in office, namely competition, firm entry and exit and technological innovations.

Unfortunately chancellors have repeatedly allowed demand to be too low such as by maintaining the fixed exchange system from 1948 to 1973 and the repeated dalliance with monetarism and the ERM fix by lawson/ major. Which is just the same old shit in a different guise. Going back further we had the return to Gold in 1925. The treasury certainly never gives up on trying to wreck the economy with depression inducing policy.

[Steven Clarke]

At the risk of being biased towards my pet policy, I suspect the biggest gains could come from liberalising the land market.

The labour and capital markets are already pretty flexible. Further attempts to liberalise them will run into diminishing returns.

[reason]

Are Japan and Italy really so pathological when you allow for demography?

[Bonnemort]

Slightly off topic, but the New York Times has discovered that an oversupply of labour DOES reduce wages - but only in Mexico.

https://www.washingtonpost.com/world/the_americas/mexico-prepares-to-absorb-a-wave-of-deportees-in-the-trump-era/2017/03/03/a7bd624a-f86c-11e6-aa1e-5f735ee31334_story.html

"More returnees means lower wages for everybody in blue-collar industries such as construction and automobile manufacturing, where competition for jobs is likely to increase, economists say."

[Bonnemort]

Damn - WaPo, not NYT. Still, they're nearly identical - almost fungible, you might say

[Tim Bassett]

My guess is that people paying too much for growth because they misprice the correlation of growth stocks. If long term growth is roughly constant then a basket of growth stocks must have negative correlation with with each other.

Thinking of growth stocks as call options shows how the overvaluation can occur. Markets are notoriously poor at dealing with correlation risk as we have seen in an almost continuous series of blow-ups over the years most notably the mortgage/CDO crisis.

Thinking about why this happens I suspect that it is because all the models I have seen take one correlation term when it is quite easy to imagine real life correlations that have both positive and negative terms plus a whole lot of noise. I have not seen any models that deal with this well and am rather unconvinced they can even exist (too many degrees of freedom to produce much useful results) but it is often useful to try and figure out boundaries in changing circumstances as different correlation factors predominate.

Eg In a normal economy two close competitors may be negatively correlated but a change in the macro environment or an innovation in the industry can rapidly switch them to exhibiting positive correlation

[C Adams]

Would it be better to say that John Landon-Lane and Peter Robertson find that similar countries with similar policies have similar growth rates?

To suggest that government policy does not have a dramatic effect on growth is crazy. Take North and South Korea.

Also small changes can have a big impact over the longer term. I would argue that government policy is very significant.