Archive for October, 2018


In October 2017 LAB had an average poll lead of 2.4% – this October Corbyn’s party is 3% behind

Monday, October 29th, 2018

The polls turned in March which coincided with Corbyn’s response to Salisbury and antisemitism becoming a big issue

October 2018 voting intention polls

October 2017  voting intention polls

Survation’s chart shows the timing of the switch


The Survation chart shows the LAB-CON splits in its Westminster voting intention polling since GE2017 when, of course, the firm was the most accurate pollster. Since then it has generally been recording the best figures for Labour and at times, like at the general election, has been out of line with other pollsters.

As can be seen things were going positively for Labour until late March when its share moved down from the 43%-45% range and has been broadly lower ever since.

There’s a great danger in looking at what was taking place at the time of the downturn and reading too much into it. Correlation is not causation.

However there were two big developments during the latter part of March – his initial response to the Salisbury chemical attack and the row about his positive comments on Facebook about a mural which was said to be antisemitic.

It was the latter that triggered the demonstration outside parliament by members of the Jewish community and their supporters as well, over the month, a series of stories about what the leader had done in the past.

We do know that in the London elections on May 4th the Labour aggregate vote was 43.9% which was way down on the 54.6% of GE2017 and all the polling for May 2018 elections.

My view is that the Salisbury response and the ongoing antisemitism row caused damage to Corbyn and his party from which it has yet to recover.

Mike Smithson


The November 6th US Midterms – where we are and what might happen

Sunday, October 28th, 2018

Viewcode looks at the detailed data

The main elections on November 6th  are for 435 seats in the House of Representatives, 35 of the 100 seats in the Senate, 36 out of 50 state/district Governors, and 3 out of 5 territory Governors. This is an attempt to summarise the predictors. The following figures were taken between 11pm BST October 27th and 3:30am BST (2:30am GMT) October 28th 2018.

Nationwide generic ballot polls

The United States is a large country with an unevenly distributed population and certain quirks on mobile phone ownership and voter registration which makes building a representative sample frame (the list from which a sample is taken) somewhat of a challenge.

Nevertheless “generic ballot” polls are conducted. A “generic ballot” poll asks the person which party they would prefer to control Congress or would vote for in their district. The website FiveThirtyEight has a page summarizing those polls and the spreadsheet is here. The last six polls so listed are here:

State level polls

Due to time constraints this will not be considered


Just like the United Kingdom, the United States has a rich ecology of analysts and academics who produce their own predictions of the outcomes. Their predictions are complicated by the concept of a “toss-up” or “too close to call” (TCTC): elements which they decline to predict, and by the use of qualifiers such as “Tilt”, “Lean”, “Likely”, etc. A summary of the predictions can be found on the Wikipedia election pages for the Senate  and House of Representatives.


Two senators (Angus King and Bernie Sanders)  are Independent but “caucus” (vote) with the Democrats. They are considered to be safe in their seats and are included with the Democrats in the table below. 35 seats are contested (24+2 Democrats, 9 Republicans). CNN has a named predictor (Harry Enten) and other predictions.

House of Representatives

All 435 seats are up for election. No independents. ). CNN has a named predictor (Harry Enten) and other predictions. The Daily Kos site is not clear enough for me to easily extract the data. 538 gives three options (lite, classic, deluxe) and two modes (one by probability, one by district). I have chosen the classic mode.

Nationwide Bets

I looked at four bookies: Ladbrokes, PaddyPower, SportingIndex (Spin), Betway. Spin were suspended at the time of inspection. The seats are deduced from the over/under. The Ladbrokes senate majority odds were deduced from Rep>50 (Rep majority), Rep=50 (no majority) and Rep<50 (Dem majority). 

State level Bets

Due to time constraints this will not be considered


I am a statistician of the old-school, based in the frequentist and mathematical traditions on limited samples. That stance is being superseded by data science which uses brute-force computer techniques on very-large samples. The latter is good for identifying outliers and perhaps better for gambling, but the former is perhaps better for providing an overview. I hope therefore that you found this article both interesting and informative.


Viewcode is a statistician who spends too much time commentating below-the-line



The persistence of lack of memory. How the state retirement age was changed and communicated

Sunday, October 28th, 2018

Old sins have long shadows. The equalisation of state pension age was first mooted in the early 1990s and was enacted in 1995. Yet it remains controversial now. The action group WASPI campaigned in the last general election and that campaign arguably made the difference in some marginals.  Theresa May might conceivably have got an overall majority if it had not been for their efforts and the whole course of Britain’s departure from the EU, among other things, might have been radically different.

What are they campaigning about? The argument shifts from point to point.  The current focus is on the absence of notice given to the women who were affected by this change. A Parliamentary Research Paper published this week, State Pension age increases for women born in the 1950s, looked at the claims on this. 

A key section has the title “Did women affected have advance warning?” Its three subsections are headed: “How much notice should people get?”; “What did the Government do?”; and “How aware were women affected?”. 

Other sources of information are dealt with in just three words (“including press coverage”). This by itself shreds the value of the report to pieces. In practice people get their information about pensions from many sources, public and private. They watch TV.  They listen to radio. They read newspapers. They look online. They look at their company pensions materials.  They plough their way through their personal pension documentation. They talk to IFAs.  They talk to friends. Any examination of the advance warning that women got needs to look at all of those sources.

As it happens, there was plenty of press coverage. I was going to look into this myself, but I found that Josephine Cumbo of the FT had beaten me to it. In the period 1993-2006, she found more than 600 mentions in the national press.  As she notes, it appeared on front pages. It appeared in tabloids. The controversial nature of the changes was fully aired. This was not a law change that was smuggled out in secret.

But let’s give the WASPI women the benefit of the doubt and accept that they somehow missed this. So let’s think carefully about what is really being said when it is complained that insufficient notice was given.  The complaint is not, on the surface at least, that the change should not have been made.  Not even a group that campaigns for the restoration of sex inequality under the name “Women Against State Pension Inequality” has the gall to argue that. The complaint is that if only I had been told, I would have done something different. 

“Something” is usually conveniently unspecified. In practice, however, it can amount to only one thing: that the individual would have saved more money earlier. Let’s leave to one side the fact that means that the money that has not been saved has in practice already been spent on something that the individual would have wanted (so the individual, having consumed cake is now wanting to have it again). This idea of earlier saving implies that the individual, if only she had been told, would have planned rigorously for her retirement.

That’s hard to reconcile with the evidence or indeed common sense. Even if the changes had been perfectly understood by all, many would not have been able to afford to do more.  Many would not have wanted to.  The tax advantages for saving in a pension had always been there, but were not taken up with anything like the enthusiasm that financial logic would have suggested.

It is disappointing, to say the least, that the Parliamentary Research Paper did not look at information about pensions provided in the private sector. In the early 1990s, occupational pension schemes were having their own torrid time equalising retirement ages.  When the state followed suit, they made sure that they built that into their scheme booklets. 

If the compilers of the Parliamentary Research Paper had done any research in this area, they would have found that the point was routinely covered in the mid-1990s in scheme booklets. Here’s a typical example (not one drafted by me, though I certainly put together a few at that time). As you can see, anyone who read such statements would have been quite clear about the nature of the change that had been made.

The conclusion we can draw about women who were members of such schemes who were unaware of the change in state pension age is that they did not look in any great detail at their pension scheme literature, or if they did, it did not sink in. That suggests a lack of interest in long term saving and casts doubt on whether they would have paid any attention at the time to any government communication either.

These, remember, were the women who were likely to be among the most interested in pensions, already having made private arrangements. The rest were going to be a lot less interested. 

As it happens, this fits well with other evidence we have. A 2005 DWP Paper on Women and Pensions noted that only 22% of women had worked out what their retirement income would be and only 47% of women said they had looked into saving or investing for retirement. Perhaps one leaflet from the government would have galvanised the indolent into action. Personally, I am seriously sceptical. 

So, contrary to the implied argument being made by WASPI and its supporters, the changes to women’s pension ages were widely discussed in the 1990s and the early 2000s and they were communicated to many by sources other than the government – and yet they still were not absorbed by many. The women affected have already had the benefit of the money that they spent and they have not noticeably been otherwise disadvantaged by any lack of notice (as opposed to disappointed).

Their claim to any form of compensation is weak. The inverse relationship between the vehemence with which they press their case and the merits of it is striking.

Upton Sinclair supposedly observed that “It is difficult to get a man to understand something when his salary depends upon his not understanding it”. It seems that it is just as difficult to get a woman to remember something when her pension depends on her not remembering it.

What MPs of all parties have to wrestle with is that WASPI and their supporters, no matter how misconceived, have votes at their disposal. Making sure that they dispose of them in a favourable manner without wrecking longstanding pensions policy positions might be a tough balancing act.

Alastair Meeks

Alastair Meeks is a former Chair of the Association of Pension Lawyers


My Cruz-O’Rourke spread bet – “selling the Republican” at 53%

Saturday, October 27th, 2018

PB regulars will know that I am a keen spread better on political events where the more you are right the more you win and the more you are wrong the more you lose. Essentially you “buy” and “sell” a position as though it was a stock or share and watch how the market moves.

At GE2017 on the Saturday after the Tory manifesto launch I “sold” CON seats at 393. They ended up with 318 MPs and my profit was the sell price minus the actual which in this case was 75 multiplied by my stake level. This was my most profitable spread bet ever.

SportingIndex have detailed spread markets available on just about every state race and I am very attracted to the Texas Senator contest where ex-presidential hopeful, Ted Cruz is trying to fight off a big challenge from Beto O’Rourke. In the previous thread David Herdson looked at O’Rourke as a presidential hopeful based on his exceptional campaign and incredible fundraising.

My bet is a “sell” of the Cruz vote share at 53% in the Senate race a week on Tuesday. If he falls short of that figure, even though he might get re-elected, I will be a winner. So if he got 51% I would make two times my stake level. If he ended up on 55% my loss would be twice the stake level.

There is a third contender in the race which will eat a bit into the Cruz-O’Rourke aggregate vote.

I was hoping to give a link to the market but SPIN tend to restrict activity in specialist areas like this to when the person running their politics markets is working. I guess he’ll be there on Monday.

I’ve been hugely impressed with O’Rourke and assuming he doesn’t bomb in the Texas Senatorial race he has a great future.

Mike Smithson


Beto O’Rourke: not the new Lincoln but perhaps following in his footsteps

Saturday, October 27th, 2018

Could losing an election be the best thing to happen to the 25/1 shot?

These are not normal times. In normal times, US presidential candidates are vice presidents, senators and governors: people who already have a record in high office. There have always been exceptions but even then, they usually conformed to the rule in a wider sense. Trump does not conform to the rule. Indeed, Trump fails to conform to many received rules of politics. The easy conclusion would be to suggest that it’s Trump who is the exception and while there’s an element of truth in that, the proper conclusion has to be that the old rules are now very imperfect guides – a conclusion strengthened by how close Sanders came to winning the nomination.*

If that’s so, what lessons should we take from it for the 2020 presidential election? The most obvious one is that we shouldn’t restrict ourselves to the usual suspects. Given that for quite a while, the Democrat field was dominated by the usual suspects – Sanders, Biden, Warren – that should have provided value elsewhere, and early backers of California Senator Kamala Harris will have already done well out of the woman who is now second favourite to win outright at 9/1 (behind Trump at 13/8).

So far, so normal. However, rising stealthily to seventh is Congressman Beto O’Rourke, priced at 25/1. For someone with only six years’ experience in the House, those would normally be extremely short odds. But as we’ve established, things are not normal. Qualifying criteria have changed and making an impression counts for more, and old-fashioned CVs, less.

    O’Rourke has benefitted hugely from that because he’s run a high-energy campaign in Texas against incumbent senator (and 2016 hopeful) Ted Cruz, raising more money in the process than any other senate candidate in history. Despite that, the likelihood is that he’ll lose.

Current polling puts him about 7% behind Cruz, which is a little more than the 3-5% it’s been for most of the campaign proper. As long as Cruz can get his vote out, he should win.

However, winning isn’t everything. No Democrat has won state-wide office in Texas since Ann Richards won the governorship in 1990 so although the state is trending back to them, to win it this time would be a remarkable achievement and O’Rourke’s result has to be seen in that context.

All the same, there are many losing candidates who run good campaigns. What makes him special? Firstly, sometimes it’s just being the right person in the right place at the right time against the right opponent. Catching the mood of the moment is something that can’t be planned, only captured, but he might have taken that chance – as his fundraising exploits have shown. Being a fresh face in a field of septugenarians also does no harm.

O’Rourke has also taken a deliberate choice to run a positive and optimistic campaign; something that stands in marked contrast to politics in general and Trump in particular. Trump might try to nickname him with something belittling – ‘the rookie’ or ‘the boy scout’ – but it’ll be hard to pin something as damaging as ‘crooked’ or ‘lying’ (though that won’t necessarily stop him from trying).

Which is where we come to a little history. The 1850s were another time of flux in America, when candidates from outside the usual circle had a better than usual chance is breaking through. One, in particular, did: a gangly lawyer from Springfield, Illinois. Lincoln was able to secure the Republican nomination in 1860 in part because his party was new and had yet to develop the full machine politics that was typical later; in part because he too had – unusually – achieved national prominence in a losing senatorial contest; and in part because as well as his own support, he was a transfer-friendly candidate. That matters as much in the days of primaries as when the convention was king.

None of which is to say that O’Rourke is a new Lincoln. It is to say, however, that Lincoln’s path to the White House is once again open. Does that make O’Rourke value at 25/1? I don’t think so. There are simply too many variables. We don’t know whether he has presidential ambitions for 2020 and while he’s run a good campaign in Texas, will that translate to the national stage? The senior members of the Democrat field are flawed and that certainly creates space for someone like O’Rourke (as in 1992, there was space for Bill Clinton to come through – though Clinton was an experienced Governor), but it also creates space for the likes of Harris and others.

In fact, the shortness of his odds is probably more a measure of the inadequacy of more prominent Democrats than of O’Rourke’s chances. I think the better strategy at this stage is to lay those front-runners.

David Herdson

* In one sense, Sanders didn’t actually come all that close. Hillary had a comfortable lead nationally during the early stages and though Sanders closed it through until about late April, when he was within a few percentage points, he could do nothing about those votes cast early on which were already in Hillary’s bag. In addition, the unpledged delegates were likely to go for the establishment candidate. However, Sanders came within a fraction of winning Iowa, did take New Hampshire and was close in a flawed Nevada contest. Had he won all three – and he could have – the national polls would likely have changed much more quickly. Further, the superdelegates might not have been so reliable for Hillary had Sanders gone into the convention with a lead in pledged delegates and votes cast.


This questionnaire on the “Values and Identity” clans of GB is well worth doing

Friday, October 26th, 2018

If you haven’t done so already check this out for yourself here.

Essentially this is a different approach to identifying voters which has been put together by the pollster BMG and academics at Bristol and Manchester.

Both David Herdson and I came out as Orange bookers.

Mike Smithson


US midterms early voting is reaching presidential election levels in some states

Friday, October 26th, 2018

But who’s benefiting?

All the signs are that turnout is going to reach almost presidential election levels in the US midterm elections that take place on Tuesday November 6th.

In many states it is possible to vote early and quite a number started the process on Tuesday with two weeks to go. Almost across the board the numbers who are availing themselves of this facility are at levels nearing, or in some cases topping. the recorded numbers casting their ballots at this point before WH2016.

This is striking because generally midterm voter turnout is substantially lower than at presidential elections. This fits in with the current polling that has enthusiasm amongst voters very high. The incumbent of the White House is a big vote driver for both sides.

Early indications are that the Republicans have edge on votes cast so far. This is based on breakdowns, available in some states, on breakdowns of early voter information that includes the party that individual register for. Whether this is significant or not is hard to say. We do know that those most inclined to vote early by post in UK elections tend to be the elderly who, of course, are much more inclined to the Tories.

The betting has seen a slight move to he Republicans for both the House and Senate majority markets. The Dems are still 65% favourite for the House.

Mike Smithson


Angela’s Ashes. Germany throws its two party system on the bonfire

Friday, October 26th, 2018

The  German state of Hessen goes to the polls this  weekend.  Hessen has been something of a swing state between the CDU and SPD. The current state assembly has the CDU in power on 38% of the vote in a coalition with the Greens  who  won 11%. The SPD leads  the opposition with 31%. But all is not well. The large national parties are worried  by the shifting ground in German  politics. A look at recent polling shows why

Both the CDU and SPD in Hessen are on track to lose a third of their vote compared to five years ago, their support is shifting to the Greens and the AfD, a move that is happening right across Germany.

In the Bavarian state election earlier this month the CSU lost control of its fiefdom as the Greens and AfD made major gains. In all the excitement of the CSU’s defeat less attention has been  paid to the SPD who suffered a much larger loss of support. So what is going on?

Germany is now entering its sixth year of the grand coalition between the CDU and SPD. The economy is booming, the government coffers are overflowing so by normal standards the voters should be giving the government credit for prosperity. Instead coalition is becoming a spiral of decline for both parties. Underlying this  are several factors

Immigration – in an act of extreme generosity or folly (take your pick) Angela Merkel  threw open Germany’s borders to refugees from Syria. Within 3 months Germany was at the centre of an immigration crisis both domestically and with its neighbours. 

Just about every bad headline Frau Merkel could have had materialised  – crises of theft, rape, murder and riots by immigrants shocked Germans. Worse, the German sense of good administration “Ordnung” was turned on its head with authorities unable to cope, bribes and corruption in the Immigration office and a total ignorance of who was entering the country.

Domestically it wasn’t pretty. With the neighbours it got worse. Angela Merkel  tried and is still trying to create a European solution to a problem Made in Germany. No one is interested; the Mediterranean countries want refugees to move onward to Germany. Eastern Europe has shut up shop. Migration is creating European tensions were previously there were none.

The crisis has cost the CDU core support among its voters and given impetus to the AfD.  All in all AfD growth mirrors the decline of the CDU, though it should be noted the SPD too has lost support among core working class voters. The migration crisis has been Merkel’s biggest mistake since entering office and now the electorate are expressing their disapproval.

Revolt in the East – Despite almost 30 years of unity the old DDR remains a place apart. Huge progress has been made in modernising the old communist state but despite it all you can’t buy optimism off the shelf. The East is still the poor cousin of the west, It has lost its future, the pied piper this time stealing  the children and taking them west so an older, but gloomier population has been left behind. Increasingly East Germans are attracted to the DDR nostalgia of Die Linke or the Germans First stance of the AfD.

Previously the CDU and SPD dominated the east, now almost one in two Ossis vote for the extremes. Centrist liberalism is in retreat and the east risks become ungovernable via the diktat of coalition arithmetic.

Demographics – The post-unification generation is now coming of age and making its mark. This is affecting politics. The older generation of war guilt and reparation is slowly dying off; its priorities and standards are disappearing with it. A similar observation can be seen in the UK with young voters caring little for Jeremy Corbyn’s past if he’s more in tune with their own concerns. 

The old certainties of the cold war two party systems are gone and with that younger voters are not afraid to ignore the taboos of the past and experiment across a wider spectrum of parties. The SPD in particular appears to be losing support among  young liberal metropolitans to the Greens.

Where’s the party? –  The two main parties have been slow to respond to the shifts taking place. Both have been jostling for space in the centre ground of politics with similar programs and common assumptions this has led to a sense of drift among supporters, coalition makes them all the same. So why vote for your old party if it no longer stands for anything?

This is a phenomenon across the West not just Germany but CDU and SPD  still hope huddling in the centre will keep them in office. Yet increasingly it is differentiation which is moving support in the country at large.

Leadership – The CDU and SPD are both experiencing a failure of leadership, but for different reasons. The SPD lost the 2017 election under Martin Schulz who refused to go in to government with Merkel. When Merkel’s coalition efforts with the FDP and Greens failed the SPD reluctantly agreed to coalition mark two. Schulz left and his successor Andrea Nahles took over to a distinct lack of enthusiasm from party members. They were perhaps right, Nahles also struggles to connect with voters  who see her as a party apparatchik. 

With the CDU the problem is Merkel‘s loss of her political sure touch, from queen of the roost to lame duck in two short years.  There is a feeling that Markel’s time has gone but as yet there is no candidate on the horizon to replace her.

So the CDU drifts on, unsure what to do next, hoping the old formulae still work and looking for something to break the stalemate. Nobody wants to wield the knife so like her doppelgänger Theresa May, Merkel still hangs on faute de mieux.

In the 1970s it was simple. Germans voted CDU or SPD and the small FDP went into coalition with the winner. In the 1980s the Greens emerged and began to pick up support. Then in the 1990s and 2000s Die Linke developed as an all-German party. Finally in this decade it has been the AfD which is new on the scene. German politics is in flux and the old monoliths are slowly crumbling.

This splintering of political opinion now makes government an exercise in traffic light management  as putting  coalitions together becomes harder and parties’ red lines are quietly dropped. The SPD are currently demanding that no-one goes  into coalition with the AfD,  ‘the heirs to Hitler’, but they dropped their own promise not to go in to government with Die Linke, the heirs  to  Stalin. Sooner or later, most likely in the East, the AfD taboo will go because the parliamentary numbers require it.

It’s hard to see a way back to previous glories for the two big parties. The coalition breaking up might help revive political fortunes. Maybe the current results could be midterm blues; the big parties have also hidden strengths such as mass membership.

But it’s quite a challenge, I suspect the big parties are dealing with a tougher problem  – Zeitgeist, the feeling that it’s time for a change, that the CDU and SPD are no longer attuned to the desires of large segments of the electorate.  It’s no longer enough to make the voters rich, you have to deliver their aspirations too.

Currently the big winners are the AfD and the Greens. The AfD stole the press headlines –  Nazi scare stories  sell – and the growth of the right attracted attention as they tore off support from Merkel’s CDU, but as PBers have noted the recent rise of the Greens has been more impressive.

If the AfD had spectacular spurts the Greens have been constant, relentless, pushing past the AfD, then the SPD and now they sit second in the polls still closing the gap with a declining CDU.

Germany now has 6 parties capable of polling over 10% in national elections; CDU, SPD, Greens, AfD, FDP , Linke. This is new territory for the Federal Republic and the rules are being made up as they go along.

If all of this seems like a far away land about which we know little then stop, go easy on the Schadenfreude.Germany is our largest trading partner and it leads the EU. Currently Europe’s largest nation is rudderless.

I often wonder if Brexit would have a different complexion if Frau Merkel was free from the daily task of keeping bickering coalition partners apart. Likewise the EU project is slowly coming to a stop from lack of leadership.

The harder question is what does this mean for the future? A multi party Germany where lego brick coalitions are put together and then broken apart is without doubt a weaker Germany. Cui bono ?

So back to Hessen. Two nervous parties are crossing their fingers it’s not a disaster for them. For the SPD it could mean the end of coalition which appears to be slowly strangling its support.For Merkel it could mean her party finally does something to move her on. Bavaria was bad, but that was the CSU not the CDU, Hessen is the home team. Those ashes from the bonfire could be her political career.