US and UK TFP since the pandemic

 The BLS has just published an update on US labour and TFP growth.  Heres a graph of  private non-farm business. Red is TFP growth, dark blue is the contribution of capital intensity, light blue is contribution of labour composition.  Note the tremendous growth in TFP, red in the last two years.  

Lets zoom into the years since 2019: 

To get the same as possible data for the UK, I used data on UK market sector value added per hour, and market sector volume of capital services.  There is no recent labour composition that i could find. The UK labour share is about 0.59.  The UK picture since 2019 is below:

What do we see?

1. in the depths of the pandemic, 2020, both countries had high labour productivity (LP) growth, but with negative TFP in the US.

2. even in 2022, both countries had negative LP and TFP growth 

3. but it is amazing that US TFP growth in 2023 and 2024 has surged.  UK TFP and LP growth is tiny or negative over those years. 

Tariffs and the UK economy

 1. There are lots of conflicting effects of tariffs on UK inflation and activity.  They are set out by Megan Greene in a very interesting recent speech.  They can nicely be summarised in two cases.

2. Case A. Unilateral tariffs. 

a.      For a given exchange rate (ER).

    i. US demand for UK exports falls. UK activity/inflation fall. 

    ii.  but, offset by foreign producers who divert cheap goods to UK.  Inflation falls. Raises real incomes, but bad for UK firms, so activity effect not clear.

b.     But the ER might change. 

                                                              i.      U US $ should appreciate, so £ depreciates relatively.  this helps UK imports to raise. good for activity.

                                                            ii.      but falling £ raises UK import prices so raises inflation. 

3. Case B.  Responses to tariffs 

    a. more tariffs everywhere raise prices.  inflation rises

b.  but more tariffs lowers demand, so inflation falls. 

c. lower demand everywhere likely has $ depreciate.  Stronger £ means lower inflation.

4. The ECB-G model gives the outcome. 

    a. output and inflation rise initially.  This is because the ECB model has a fast-moving ER channel that dominates, so £ depreciates and there is trade diversion.  Thus import prices rise and so inflation rises.  There is also trade diversion, which lowers inflation and monetary policy reacts.

Now, the 

5. now consider the case of a response.  here all countries respond, activity falls and the US $ likely falls. 

Now we get the opposite effect

with falling output and inflation. 

6. what might happen to adjust these scenarios?  Megan considers: 

a. supply chain disruptions might lower feasible supply, pushing up on inflation

b. a flight to safety might make the $ appreciate.  This is important in the ECB model which she says is dominated by the exchange rate

c. monetary policy is endogenous and passes quickly through to inflation and output.

To summarise, my understanding of this is that it looks like:

a. the SR effects in this model are driven by ERs and a bit of trade diversion.  The longer term effects by monetary policy. 

b. in both cases, trade diversion lowers inflation and is uneven for activity.  Weaker demand lowers both.  

c. But with unilateral, $ appreciates and retailation, $ depreciates.  

There is a nice table summarising

with the ER movement row the crucial one. 

My comment would be

1. the longer run effects of supply are neutral in most models

2. the effects on goods import prices work thru most models quickly but take a time to play out. 

3. the Bank is limited in what it can do on lowering rates since core inflation is high.  If that remains high, then high goods inflation will cause the inflation target to be overshot even more.  Low goods inflation will hide this underlying service sector problem.  That low goods inflation seems likely to be fed by trade diversion.  

4. Broadly, so far this year the $ has got weaker and the £ stronger.  This bears down on UK inflation. 

How to reduce inequality: ban the National Lottery and Nobel Prizes

 The excellent and insightful Nick Oulton (https://www.lse.ac.uk/CFM/assets/pdf/CFM-Discussion-Papers-2022/CFMDP2022-05-Paper.pdf), link  has a very interesting thought.

1. People often don't want inequality. 

2. But an aversion to inequality likely comes with some moral attitudes.  One example is whether the inequality is deserved or not.  As he says if we didn't care about "deservedness"...

then social welfare would be raised by abolishing two institutions (among others): the national lotteries run in many countries and the Nobel prizes. Both increase inequality unambiguously. Indeed, Nobel prizes must be the most unequally distributed of all forms of income: only a dozen or so individuals receive one each year out of a world population of some 8 billion.

 

Nobel prizes could be justified on Rawlsian grounds: monetary incentives are needed to induce the effort required to make discoveries that benefit everyone, including the worst off. But suppose that it could be conclusively shown that the monetary rewards are not necessary, and that the prize winners (and their less-successful colleagues) would have expended the same effort in exchange for just the honour and glory alone? I suspect that most people would still be quite happy to see the winners receive a monetary reward, even if it was not economically required. This is because they are perceived to deserve it.

 

With national lotteries, a different form of desert comes into play. In the UK version, some winners receive £20 million or more, and, in one sense, no-one is worth this amount. But anyone can buy a lottery ticket and, as long as the lottery process is perceived as fair (not rigged), most people are quite happy with the outcome.

What is a log point? 100*ln(new/old).

 Nerdy. Often when doing growth in Economics, we use change in natural logs.  For a change to y from x, the log point change = 100*ln(y/x).   So a change of 1 in the natural logs, which we often call "1%" is 100 log points.  

If a dataseries rises from 100 to 100.5, then: 

a. the % change is 0.5%

b. the change in the natural log is 0.0049875

c. the change in log points is 0.49875

If a dataseries rises from 100 to 101, then: 

a. the % change is 1%

b. the change in the natural log is 0.00995

c. the change in log points is 0.995

A basis point is defined as: 1bp is 0.0001 = 1/100th of 1%.  Or 100bps are 0.01 = 1%.  One might be tempted to say 0.995 log points are 99.5 basis points, but that's not often done.