Why Keynesians are wrong

The most prominent theory in macroeconomics is New Keynesianism. One of the most striking and unique predictions that New Keynesianism makes is that when the economy is in a recession, everything gets flipped upside down. Specifically, when interest rates are at the zero lower bound and the economy is stuck in a liquidity trap, most of the things that would usually improve economic outcomes actually worsen them.

The NK model predicts that supply-side loosenings, like lifting employment regulations, cutting taxes or liberalising immigration laws, will actually make things worse in a recession, as will interventions that increase price flexibility. However, this prediction—familiar from Paul Krugman's NYT columns since 2007—seems to have been strongly challenged in a batch of recent papers.

The first is "Supply-Side Policies in the Depression: Evidence from France", by Jérémie Cohen-Setton, Joshua K. Hausman, and Johannes F. Wieland. It, as the title suggests, looks at data in from the great depression in France, one of the areas that suffered it from the longest, due to the obsessive desire of the Bank of France never to sever the currency's link with gold. The Keynesian model would predict that devaluation and leaving gold were the only game in town, but in fact the negative supply-side shocks that happened at the same time depressed activity, even in a deep slump.

The effects of supply-side policies in depressed economies are controversial. We shed light on this debate using evidence from France in the 1930s. In 1936, France departed from the gold standard and implemented mandatory wage increases and hours restrictions. Deflation ended but output stagnated. We present time-series and cross-sectional evidence that these supply-side policies, in particular the 40-hour law, contributed to French stagflation. These results are inconsistent both with the standard one-sector new Keynesian model and with a medium scale, multi-sector model calibrated to match our cross-sectional estimates. We conclude that the new Keynesian model is a poor guide to the effects of supply-side shocks in depressed economies.

The second is "Are Supply Shocks Contractionary at the ZLB? Evidence from Utilization-Adjusted TFP Data", by Julio Garín, Robert Lester, and Eric Sims. It looks at more extensive data on productivity. The Keynesian model predicts worse productivity improvements from supply shocks that occur in slumps but the data finds quite the opposite result.

The basic New Keynesian model predicts that positive supply shocks are less expansionary at the zero lower bound (ZLB) compared to periods of active monetary policy. We test this prediction empirically using Fernald's (2014) utilization-adjusted total factor productivity series, which we take as a measure of exogenous productivity. In contrast to the predictions of the model, positive productivity shocks are estimated to be more expansionary at the ZLB compared to normal times. However, in line with the predictions of the basic model, positive productivity shocks have a stronger negative effect on inflation at the ZLB.

The third, "What Was Bad for General Motors Was Bad for America: The Automobile Industry and the 1937/38 Recession" by Joshua K. Hausman, tackles the question less directly, finding that shocks that impacted the car industry, even if they weren't aggregate, demand-side shocks, nevertheless had large impacts on overall output and income.

I think the New Keynesian model is wrong about a lot of things. It seems that the impact of supply-side moves in a recession is yet another prediction it gets wrong.


This is a referendum on the EU, not the single market

Today three opinion polls - from YouGov, ICM and TNS - are all putting Leave in the lead. Those are on top of other very recent surveys suggesting a shift in Leave’s direction.

That has obvious significance and it is causing reverberations in the Remain camp.

Taking a step back for a moment to last week, on Tuesday we saw the Telegraph’s Ambrose Evans-Pritchard coming out in favour of the Flexcit plan by Dr Richard North, and my own ASI paper called “The Case for the EEA Option” that borrows from the North plan.

That article prompted leading Remain thinker Charles Grant of the Centre for European Reform to agree that using the EEA as a transition point was indeed a viable exit option - to my knowledge, the first Remainer to do so. Grant also agreed that parliamentary arithmetic very much favoured this option and, further, that top Vote Leave MPs would be able to support such a manoeuvre after a Leave vote.

Now this morning, the BBC’s James Landale has reported that Remain MPs are indeed saying that while they would have to deliver a Leave proposition after a Leave vote, they would not support leaving the single market (the European Economic Area), which is separate to the EU.

There are two ways to view this. Firstly that they are genuinely concerned about leaving the single market more than leaving the EU. That would make perfect sense as so many of their objections to leaving the EU are actually objections to leaving the single market. But secondly that they are making mischief for the Vote Leave campaign that has nailed its colours so firmly to the border control mast.

I suspect it is the first but with a helpful side-effect (to the Remain camp) of the second.

And of course the other paradoxical side-effect is that it derisks the economics of Leave and makes Leave more attractive to wavering voters.

Sure enough, Dominic Cummings for Vote Leave responded by suggesting that MPs were saying they would ignore a Leave result. Yet that is exactly what they are not saying. Rather that “Leave only means leaving EU membership”. The BBC report made it very clear:

One minister said: ‘This is not fantasy. This is a huge probability. The longer we move away from the referendum, the more the economic pressures will grow to keep some links with the single market.’
Another said: ‘We would accept the mandate of the people to leave the EU.’

Indeed they would. Because in the event of a Leave vote, the precise question on the referendum ballot will matter. A lot.

Stop the facade, it's time to come clean about national insurance contributions

Tax Freedom Day, which measures the total tax burden compared to Brits' total incomes, falls on June 3rd this year. Everything we earn from every source for 153 days goes to the government and its programs; everything from today to the end of the year we spend ourselves. This is the latest it has been for fifteen years, and the evidence suggests a big government is an unwieldy one, and one that reduces growth. One way we could reduce it in a way that the average person will really feel is by raising the threshold you have to earn before you pay national insurance.

Britain used to have a contributory welfare system, driven by the recommendations in the Beveridge report, a surprise best-seller in 1942. People paid in, through national insurance, and gained eligibility for various social insurance and welfare benefits: pensions, the dole, sickness benefits, and so on. 

We still pay national insurance contributions, fooling many into believing that there is still a pot in which these pile up, ready to pay for our needs when we age, or lose our jobs. This is a façade; the pot no longer exists. The contributory system has effectively withered away, and we should cut some of the complexity out of our tax system by making that official.

People pay national insurance if they earn above £155 a week (equivalent to £8,060 a year). The fact that eligibility is determined by weekly pay, rather than annual salary, hearkens back to the system's 1940s origin. But these contributions have become tied only nominally to the receipt of benefits.

The last government did away with most of the complexities in the state pension to do with different levels of payment. That means it is all-but universally given at the same level now: £155.65 per week, and will in principle be given only to those with 35 years of national insurance contributions. But the current system gives national insurance credits not only to those earning above £155.65, but also those claiming jobseeker's allowance, employment and support allowance, child benefit for kids under 12, or carer's allowance, as well as those earning above £112, and so not paying.

The government, thankfully, is no longer trying to do the job of pension providers. It is trying to guarantee that no pensioners need to live in poverty. In so doing the contributory link has been broken on both payments; you don't need to pay to generate eligibility and the amount you pay does not affect the size of the payments you're eligible to. But because people are told they're contributing, some still believe they are saving up for their retirement, or for difficult spells, when they hand over NICs to the government.

We should acknowledge this disconnect by rolling national insurance contributions into the income tax, and starting them at the same threshold, so people know exactly how much they're paying. What's more, this should be our main strategy for tackling low pay, rather than the national living wage. The Adam Smith Institute’s recent paper 'Abolish the Poor' showed how simply lifting those working full time on the minimum wage out of income tax and NICs would bring them to the living wage income, without the risks of unemployment or an early shift to automation that come with wage price fixing.

The government has already made impressive steps on income tax, taking many low paid workers out of the levy entirely, but NICs have been left by the wayside. An NIC cut would boost work incentives and return money to the badly off at the same time, without risking unemployment.

The link between contributions and benefit receipt is gone in all but name. It was a system built for a time when a single earner was expected to support a family, and where the biggest problem was unemployment, not low wages. In practice, the welfare system has adjusted to the new reality, but the language and framework remain. We should recognise the situation as it is and cut away the final remnants of the old contributory link, rolling NICs into income tax and raising the NIC threshold so that we are taking less tax from the lowest earners. Doing so could give minimum wage workers a living wage, without the unemployment that risks. Beveridge's ideas may have been good for 1942, but they are not a good fit in 2016.

We rather disagree with this estimation of Bitcoin

We're rather techno-optimists around here: largely because we know a bit of economic history. 250 years back the average human was able to consume some $3 a day of current value. Today that's more like $30 and we in the rich countries are doing very much better than that too. All driven by technological advance: sure, that itself incentivised by markets, economic freedom, the price system and getting the institutions right. But it is technological advance which is the direct cause of our current wealth.

However, we're rather less optimistic about the prospects for Bitcoin. We're sure that the blockchain will be used to do some interesting things, although pretty certain that it will be a less clunky version of it which is. Bitcoin itself we don't think is going that far. And we're really pretty sure that it's not about to eviscerate banking:  

New technology such as artificial intelligence and blockchainwill utterly shake up the fundamental principles of banking, challenging the entire industry according to former Barclays chief Antony Jenkins.

He believes the innovation in finance could eliminate the need for maturity transformation – the process by which short-term deposits, such as current accounts and instant access savings, fund long-term loans including mortgages.

That is a fundamental principle of the industry as banks can offer a low interest rate to savers while charging more to borrowers, profiting from the gap between the two rates. Yet in 10 to 20 years’ time, he believes the need for banks to perform the function might no longer exist – already some investors are sidestepping banks by using websites to match borrowers and savers directly.

That there will be peer to peer lending we have no doubt about. But look at what the prediction really is: Bitcoin, or alt-currencies, or the blockchain, will wipe out fractional reserve banking. We will end up with a system instead of only 100% reserve banking. That's what no maturity transformation means. And we really do not think that is going to happen.

Simply because that maturity transformation is too damn useful. The desired maturity and liquidity of savings is rather lower than the desired such of borrowers. Thus, somewhere in the system we need maturity transformation. And as Brad Delong likes to point out if you borrow short and lend long then you are a bank: that being the definition of what banking is. Further, if you're not, you're not doing banking.

The blockchain might have all sorts of fun uses but it's not going to replace that basic desire we have for maturity transformation. Thus it's not going to replace banking.

It's official, Uber saves lives

Tomorrow, the European Commission is set release its findings from a year long investigation into the Sharing Economy. Interestingly they're expected to go against many European cities who are increasingly trying to regulate ridesharing out of existence.

Last week, I argued that attempts to protect consumers by regulating supposedly unsafe products often backfired, because those products are often substitutes for even riskier products and behaviours. In particular, I pointed to Austin's fingerprint background check requirement that made Uber and Lyft prohibitively expensive to run in the city, and led to consumers taking even riskier journeys. But, if anything the harms of driving Uber and Lyft out of town are even greater than I made out.

At least that's according to a new Working Paper from Angela Dills and Sean Mulholland, which investigated the impact of Uber's entry to a city on vehicle accidents and crime. They looked at data across 150 cities and counties on a range of metrics, from vehicle accidents and DUI arrests, to drunk and disorderly conduct and aggravated assault. 

Dills and Mulholland point out that there's a range of ways that Uber might affect crime and accident rates. Maybe it'll increase accidents because Uber drivers are more likely to be distracted both by smart phones and passengers. Maybe it'll increase driver-on-passenger assaults as unsafe drivers sneak through lax background checks. Maybe it'll increase assault widely, as more folks get their drink on, safe in the knowledge they don't have to be the designated driver.

But the data doesn't bear out any of these fears. Dills and Mulholland came away with two major findings.

First, the rate of vehicular accidents falls quite dramatically when Uber enters a city, with traffic fatalities declining by 16.6 per cent over a year. This can be explained by both a reduction in the number of people driving under the influence, as well as the fact that the people most likely to use Uber (i.e. millennials) are terrible drivers and anything that keeps them off the road is a good thing.

Second, they find declines in arrests for both assaults and disorderly conduct. This may be because Uber reduces passenger wait times, lowering the risk of someone being attacked while waiting for a cab. This finding is especially important as governments have attempted to impose minimum wait times on ridesharing services with varying success (thankfully, TFL's proposals were roundly rejected).

As city governments increasingly move to crack down upon ride-sharing services like Uber on the grounds of public safety, legislators should take these findings very seriously. Bans on Uber aren't just bad economics, they can kill.