Brexit and Hate Crime

After the UK voted to leave the EU it was widely reported that there was an increase in hate crime. I was initially fairly sceptical of this and thought that a large proportion of this increase was likely to be explained by individuals being more likely to report crimes committed against them as hate crimes. For example if individuals believed that the Brexit campaign involved significant levels of xenophobic rhetoric they might be more likely to interpret crimes committed against them as being racially or religiously motivated. This effect would likely be compounded by reporting programmes like True Vision. In addition, the hate crime increase appeared to affect LGBT people despite the Brexit campaign barely touching on LGBT issues – leaving no clear causal link between Brexit and the increase in hate crimes committed against LGBT individuals.

The chart below shows quarter on quarter changes in hate crime across England and Wales. There appears to be some degree of seasonality. As a result even though there was an increase in hate crime in the quarter leading up to and including the period immediately after the vote (Q2 – April-June), this appears unremarkable compared to the similar quarter on quarter changes in previous years. What does stand out is the period after that (Q3 – July-September), where the increase in hate crime was larger than usual.

brexit chart

 The task of disentangling whether the increase was real or driven by reporting is a non-trivial one – at first glance both might equally explain the change.  If we assume that some small but fixed proportion of Brexit supporters committed these crimes then we would expect to see regional variation in the increase in hate crime related to the proportion of people voting for Brexit. If the change was driven by reporting we would probably expect to see a more uniform change across the country (although might still see some effect, if people in predominantly Leave supporting areas were more likely to be worried about hate crime and flag crimes within this category or if police forces tended to prompt individuals in these regions more about potential hate crime links).

The police crime and outcomes data are reported on the basis of Police Authorities, which can then be used as a basis to aggregate the Brexit referendum results data. All crimes which are listed as “religiously or racially aggravated” are tagged as hate crimes. The City of London is excluded from this analysis as the profile of the small number of residents who live there likely differs substantially from that of the individuals who work or visit the square mile. Running a regression of the Q3 proportional increase in hate crime on the proportion voting Brexit produces the following results:

HATECRIME16Q3 = -0.279 + 1.061*LEAVE
se:                              (0.387)   (0.696)

The relationship is statistically insignificant (p=0.14) although this isn’t entirely surprising given the low number of observations able to be used for this analysis (n=42). Furthermore the proportion voting Brexit only appears to explain around 6% of variation in hate crime – which suggests that this is not an overwhelming driver. Although this doesn’t meet conventional levels of statistical significance these results do at least suggest that there is more likely to be a correlation than not.

There are however a few major cautions in believing even this limited story. First if you look at the quarter leading up to Brexit when most of the supposedly “toxic” campaign rhetoric emerged there is no correlation between the increase in hate crime and the regions that voted leave:

HATECRIME16Q2 = 0.391 – 0.235*LEAVE
se:                              (0.370)  (0.665)

Secondly if you look at the quarter-on-quarter increase in hate crime in Q3 there is a much more substantial correlation with voting for Brexit for 2015 than 2016, when the campaign actually happened. This result is statistically significant (p=0.014) and explains a greater proportion of variation than the subsequent year (R^2 = 0.142 as compared with R^2 = 0.055). As a result it may be the case that there are other reasons for variation in hate crime that happen to correlate with voting Leave but are unrelated to the Brexit vote itself:

HATECRIME15Q3 = -0.632 + 1.277*LEAVE
se:                               (0.276)   (0.497)

This is clearly insufficient to give a definitive answer on what caused the increase in hate crime. However it suggests that at the very least the story is a lot more complicated than the one that we have been left to infer – that Brexit supporters whipped up by xenophobic rhetoric went on to commit these crimes. There may of course be reasons why those most likely to commit hate crimes were more evenly distributed than Leave voters – although these don’t seem immediately obvious.


How “Random” are Debating Results?

Randomness can mean a number of things. In this post I use it in a very particular sense. “Randomness” is the extent to which Team B is likely to beat Team A in spite of the fact that Team A is “better” ie. would win the majority of times against Team B if maintaining their present levels of ability across an infinite number of debates. This can be thought of using a simple model:

performance = latent ability + shock

Team A beats Team B if its performance score is higher in a given debate. This is a function of both the actual ability of the teams and a shock which reflects the degree of “randomness”. Making distributional assumptions allows us to analyse the scale of this shock. I take the distribution of abilities as well as the shocks to be normally distributed, although alternative specifications are of course possible and worth exploring. There are no natural units to use here and a doubling of the variance of the ability distribution would be observationally equivalent to a halving of the variance of the distribution of the shock. As a result identification requires fixing one of these variances. I take latent ability to be drawn from N(0,1) and the shock to be drawn from N(0,s) where s is an unknown parameter. There are some issues with this. Draws are independent – this may be problematic if confidence or determination depend significantly on previous results. Shocks may also be negatively serially correlated at the team level if driven by relative strengths in particular positions, given positional rotation. Furthermore the presence of some very good teams may allow teams of similar but lesser ability to converge towards the better team. These are all difficult to model and in any case do not strike me as massive concerns, so are ignored for now.

Determining s is not straightforward. We cannot simply look at the number of “upsets” that occur. We might consider a debate an upset if a team that finishes higher placed upon team tab lost to a lower ranked team. The problem with this is we only observe the final positions and not the latent abilities. In an environment with randomness the team that is lower on team tab may well be the better one and so that debate wouldn’t be a tue upset at all. However a particular feature of power-pairing allows estimation. In a world where s=0 and the results of a given room were fully deterministic the top team at an n round competition would always finish on 3n points. Hence at WUDC the top team would always break on 27 points. Given sufficient rounds, as the level of randomness increases the average number of points the top team on team tab will finish on decreases.

I looked at the approximate number of teams at the last 10 WUDCs and simulated 10,000 tournaments of each size for a number of values for s, given the assumptions above. Performance corresponds to speaks and the relative position of speaks within a room determine team points. Allocations are randomised for round 1 and power paired thereafter with randomised pull-ups. This took more than a day of calculations on my three and a half year old laptop. I suspect this could be done much more efficiently but is in any case fairly computationally intensive. When s=0.75 and the number of teams is 380 there were 486 instances where the top breaking team finished on +9, 1100 on +8, 1909 on +7, 3052 on +6, 3072 on +5 and 381 on +4. These estimated proportions allow the approximation of a likelihood function for different values of s given the actual number of points the top breaking team had in each of these ten years. This is obviously a tiny amount of data but is intended to illustrate the approach. Other teams beyond the top breaking could be added although I’m unsure whether these will be as informative. If we only gain one observed value for each WUDC our sample size will clearly be highly limited. Many other competitions could be incorporated although we could probably expect a higher variance of results relative to the variance of team abilities on regional circuits if the abilities of teams are correlated with geography, meaning that s could differ. The likelihood function using my limited dataset is shown below. The function reaches a maximum when s is a bit less than 1, although tails off slowly.

likelihood function

What does this mean in practice? If s=0.95 and WUDC has 360 teams we would expect the top breaking team to finish on 27 around 3% of the time. They would break on 24 around 31% of the time and on 23 around 36% of the time. We would also expect the team that breaks first to be the highest ability team at the competition around 44% of the time. These exact figures may be sensitive to assumptions about the distribution of the shock but likely provide a useful guide to comparing different tournament designs.

Height, Nutrition and Living Standards in Nineteenth Century New Zealand

In the aftermath of WWII heights were almost identical in the North and South of the ethnically homogeneous Korean Peninsula. Today the divergence between the capitalist South and the centrally planned North can not only be seen from space but also in the 3-8cm difference in heights. Average height is a function of the net nutritional status of a population during childhood and adolescence. This not only reflects the quality and quantity of food consumed (which historically constituted a very large proportion of expenditure) but also the energy the body expends in warding off disease or labouring. Given diminishing returns to height in nutrition, average height will also reflect distributional concerns. As a result anthropometric data has been used in a wide variety of historical contexts, most notably in contributing to the debate about the longest running controversy in economic history – the question over what happened to the standard of living of British workers during the Industrial Revolution.

Height data can be especially illuminating for periods where we lack good quality national income data. For example New Zealand GDP data for the nineteenth century is calculated indirectly using the equation of exchange using the money supply, price level and velocity. The latter is determined based on a cointegrating vector between Australian and New Zealand aggregates on the basis of other periods. Given the potential for error in this method, Angus Maddison describes the retrospective estimates that exist for New Zealand as the among the weakest of all OECD countries. Two existing studies by the same set of co-authors analyse height in nineteenth century New Zealand. The first analyses WWI records and finds no statistically significant difference in average height between those born in the 1880s and those born previously. No regional differences are found. The second uses all available prison registers and small sample from the Police Gazettes from 1872-1892. This dataset relies heavily on Napier and New Plymouth registers as the records of the major population centres were never transferred to Archives New Zealand. While there is evidence for an increase in height in the 1890s there is less variation in the preceding period with decadal fixed effects for the period from the 1850s-1880s lying within around 0.3cm of each other. Before 1900 there was little difference in Maori and Pakeha heights once occupational controls are introduced.


I extracted and analysed a new dataset from the New Zealand Police Gazettes. These newsletters were published weekly and distributed to police stations from 1861-1990. As well as containing notices about the day-to-day work of the police they contain tabulated records of those released from prison. The data available includes the name of the offender, jail, place of trial, offences committed, year of birth, place of birth, trade and height.

police gaze.jpg

My dataset includes 6420 discharge records from 1892-1910 of criminals born in NZ. An automated matching process to remove multiple discharges of the same person leaves 5599 people. Restricting the sample to those aged over 21 (to remove potential sample composition bias if individuals are still growing) there are 3673 men and 524 women.

Headline Results

Unsurprisingly my dataset indicates that height in New Zealand compared favourably to many of the Western European nations contained in the Hatton-Bray dataset. This is likely a reflection of the diet of nineteenth century New Zealanders which often involved eating meat for three meals a day (bacon/chops/offal for breakfast, cold meat/sausages for lunch and roast/stews for dinner).


The chart below shows how male height evolved from 1850-1890 (5 year centred MA). Although living standards appear to be high by international standards there was little improvement in height over the period and substantial cyclical variation. Decadal aggregation suggests no significant changes across the entire period, however one period stands out if 5 year breakdowns are used, with those born from 1875-79 around 0.6cm shorter  than those born from 1850-1869. Although it’s difficult to disentangle the exact causes as height can be affected by shocks throughout the whole of childhood and adolescence and there exist complex interaction/catch-up effects this coincides with a cohort we would expect to be affected by the depression following the Vogel boom but who would not have benefited from rising export prices in the mid-1890s.


Regional Variation

Assuming a correlation between place of birth and place of trial it is also possible to investigate regional differences in living standards. The data suggest the existence of an urban height penalty of around 1cm, although surprisingly this appears to apply fairly evenly to towns with a population of above 3000 and not just to the very largest commercial and industrial centres such as Auckland and Dunedin. One possible explanation is that this gap represents the greater availability of protein in rural areas rather than disamenities in urban areas. However contemporary accounts suggest substantial hazards of urban living. The Otago Daily Times wrote in 1864 that Dunedin had “reproduced with faithful accuracy the wretched tenements and filthy back slums of an English town”. An 1870 editorial of the Auckland Star reported on an instance of “Asiatic cholera” and opined “That miasmic poisoning is slowly undermining the physical stamina of citizens must be a conviction in the mind of everyone that has breathed here”. Dunedin was the nation’s main industrial hub and the city’s constituent boroughs were largely unwilling to enter into coordinated sewerage projects until 1876. Auckland did not develop a comprehensive sewerage and drainage system until the early 20th century and until then most people relied on nightly collection by soil carts and what little did exist depended heavily on a partially open channel on Queen St that discharged into the Waitemata Harbour. The data do not readily allow us to distinguish between these hypotheses but the econometric evidence for the urban height penalty is strong. There is also some evidence for a North Island height penalty which would reflect the period being one of “Middle Island Ascendancy”.

Regression on subsamples for each major centre reveals one major additional finding. Those born in Auckland from 1880-84 were some 2cm taller than those born 1850-1870, with this height premium even heigher when compared with those born from 1875-79 and 1885-89. This variation is greater than both the time invariant regional differences and the temporal differences and is hence of particular interest. One explanation may lie in the weak state of the Auckland economy in the late 1860s, which would depress the comparison height. The movement of the capital to Wellington in 1865 resulted in a substantial loss of economic power. This was coupled with the withdrawal of the imperial troops who had been fighting the Waikato campaigns of the New Zealand Wars and led to significant unemployment. However this explanation is incomplete as it would also affect other periods. A more likely explanation can be found in the account of Russell Stone. He argues that Auckland emerged from depression earlier than the rest of New Zealand with significant commercial growth in the early 1880s. This had come to an end by 1888 when the BNZ was forced to reveal significant losses on its Auckland based lending and subsequently restricted credit resulting in depression. Gary Hawke subsequently questioned whether contemporary accounts of “depression” were sufficient to make this conclusion given the ambiguity of the term as used at the time. Although GDP data would be needed to answer this question my height data appears consistent with Stone’s narrative given the reversion in the height increase for birth cohorts from 1885-89, that would not seem to merely be a slow-down in growth but something more significant.

Sample Males 21+ Males 21+ Males 21+; Auckland
Constant 171.4*** (0.334) 171.96*** (0.358) 169.48*** (0.458)
Born 1870-74 0.247 (0.324) 0.266 (0.324) 0.805 (0.722)
Born 1875-79 -0.622** (0.230) -0.586** (0.299) -0.426 (0.690)
Born 1880-84 0.016 (0.319) 0.046 (0.319) 2.041*** (0.740)
Born 1885-89 -0.109 (0.388) -0.092 (0.388) -0.281 (0.840)
Auckland -1.095*** (0.303) -0.256 (0.404)
Wellington -0.924*** (0.343) -0.087 (0.435)
Christchurch -0.279 (0.438)
Dunedin -1.087** (0.479) -0.646 (0.455)
North Island -0.333 (0.336) -0.729** (0.329)
Major Town -1.031*** (0.329)

* significant at 10% level; ** significant at 5% level; *** significant at 1% level
Estimation by OLS; Standard errors shown in brackets


Existing work in into NZ height in the nineteenth century ignores female height. This is understandable as most female prison registers are missing, women consisted a much lower proportion of the prison population  and women did not feature in military data. In addition it is not possible to run the same checks for the robustness of results as it was possible to for men due to of a lack of occupations given in the data. The small sample size limits the usefulness of robustness analysis by crime committed as I was able to do with men. Moreover a large proportion of women in the sample were arrested for prostitution and may hence be less representative of working class New Zealanders than the male sample, which consists of a substantial proportion of petty and low level crime. With these reservations in mind, there does appear to be a clear upward trend in female height.


These increases are large and statistically significant even given the substantial standard errors, with point estimates of over 3.5cm for birth cohorts from 1885-89 as compared with 1850-69. The finding of an urban height penalty applies to women as well as men.

Sample Females 21+ Females 21+; excl prostitutes
Constant 159.34*** (1.347) 159.41*** (1.928)
Born 1870-74 2.415** (1.103) 0.517 (1.841)
Born 1875-79 2.300** (1.039) 1.418 (1.691)
Born 1880-84 2.617** (1.105) 3.595** (1.758)
Born 1885-89 3.823*** (1.419) 5.034** (2.245)
North Island 0.353 (0.792) 0.009 (1.257)
Major Town -2.322** (1.143) -1.919 (1.585)

* significant at 10% level; ** significant at 5% level; *** significant at 1% level
Estimation by OLS; Standard errors shown in brackets

One possible explanation can be found in the high numbers of men of marriageable age relative to women in nineteenth century New Zealand. An optimising household aiming to maximise the utility of its members might invest greater resources in its daughters in such a situation as a means of intergenerational social advancement given their strong marriage prospects. The decision of optimal resource allocation might follow some kind of learning process such that over this 40 year period the nutritional standards of girls approached this higher level. Although caution is needed in interpreting this result, the magnitude of the effect size is striking and worthy of further analysis.

NZ Flag Results Analysis

The results are in with around 57% of voters choosing to keep the current flag and 43% choosing the Kyle Lockwood Silver Fern design. I am personally disappointed by this result but it ended up being a fair bit closer than I had expected. The campaign has been especially odd with most of the “debate” being focused around whether this is a John Key vanity project and whether the $26m could have been spent better some other way, rather than which flag represents us best as a country. I decided to have a very quick look at the preliminary voting data from the Electoral Commission. This data is broken down by electorate, so using the characteristics of electorates it is possible to get some sense of what might have been driving decision making (although obviously the true story is complex and multi-causal).

One recent poll showed much greater support for the current flag amongst young people. The graph below plots the percentage of support for the new flag on the y-axis and the median age of individuals in each electorate on the general roll on the x-axis. There does indeed seem to be some positive relationship between age and voting for change. Intuitively this seems surprising as we might expect those of older generations to have a greater attachment to the traditional flag and the symbolism of ties to the United Kingdom. Obviously the measure used below is imperfect as the age distribution might affect support within a given electorate without necessarily affecting the median – in principle it is possible to use more detailed demographic breakdowns but for simplicity I stuck to this simple measure.


There also appears to be some indication of a positive correlation with median income in an electorate and supporting the fern.


However interestingly enough there seems to be another factor which acts as an even stronger predictor of support for the fern – namely the proportion of the party vote received in the 2014 election by the National Party.


This final factor’s explanatory power is robust to controlling for the first two. The estimated equation is given below with standard errors in brackets.

FERN =     3.08       +     0.414*NAT      +       1.796*LOGINCOME       +         0.0035*AGE
s.e.           (20.24)             (0.0359)                    (1.704)                                          (0.0863)


The coefficient on NAT is statistically significant at a 1% level but neither of the other two factors are. As a result it may be the case that the apparent association of young people with the old flag is an artifact of a correlation between youth and not supporting National. Obviously there are many more factors uncontrolled for which could be correlated with supporting National and voting for the new flag – but the strength of the correlation and the apparently weak explanatory power of the other factors has me wondering whether New Zealand missed a chance to have a proper discussion on a change to how we present our identity as a result of party politics.


You Won’t Believe These 5 Times Peter Leeson Applied Rational Choice Frameworks To Unusual Topics

Milton Friedman famously owned a car with the number plate MVPQ. However as far as we know he never went so far as to get an economics tattoo. One man who did take that next step was Peter Leeson, a professor at George Mason University, who marked his biceps with supply and demand curves aged 17. Here are 5 reasons Peter Leeson is one of the niftiest economists writing today.

To the Economicsmobile!

1) Trials by Ordeal

Most people look at trials by ordeal and dismiss them as a primitive superstition. Not Peter Leeson! He argues that ordeals created a separating equilibrium to correctly distinguish between the guilty and innocent. If the guilty feared conviction by ordeal they would be unwilling to risk going through with the ordeal but the innocent believed God would exonerate them. Given this separation the priest could rig the trial and correctly find the accused innocent, upon observing their willingness to go through with the ordeal. This explains why ordeals were not used for non-believers and why most accused criminals succeeded in apparently impossible tasks!


Ordeals: superstition or separating equilibrium?

2) Human Sacrifice

Human sacrifice seems so brutish and horrific that it couldn’t possibly be explained by the decisions of rational utility maximising agents. Or could it? Imagine you are a tribe fearing attack by your neighbours. In order to avoid attack you need to convince your enemies that you are not worth raiding. You could demonstrate your destruction of wealth by burning crops but this could be faked – you could burn a small layer of valuable crops that cover up low value waste. The logical solution? Buy slaves at high prices from others and kill them in a huge festival to demonstrate that you aren’t worth robbing!


Defending property rights?

3) Pirates

Despite being lawless rogues pirates developed highly sophisticated systems of governance. Pirate codes acted as constitutions and constrained the power of captains. What’s more interesting is that the distribution of booty was often highly equal, with captains frequently only getting twice as much as the lowest ranked pirates as compared with four to six times as much in the merchant marine. This helped to avoid costly potential conflict on board and ensured a greater degree of homogeneity in desired retirement timing, leading to higher effort in battle by crews.



4) Longbows

For more than 150 years the English had a monopoly on the use of the longbow in war. Longbows were cheaper to manufacture than crossbows but relied upon large numbers of trained archers to be successful. English kings banned activities that might compete with archery training, built up a culture around shooting and even required merchants to import a given number of bow staves on each shipment. Why didn’t France and Scotland do this? They were too politically unstable and so their rulers chose to avoid adopting a weapon ideal for use in a rebellion against them.

District 12: Proof the longbow can still emerge with weak institutions

District 12: A tragic combination of a low factor endowments and weak, extractive institutions

5) Animal Trials

The medieval church tried weevils, grasshoppers and rats as legal persons in France, Italy and Switzerland. Were they crazy? Nope, Leeson answers, just doing as any other rational profit-maximising agent would. The church had an incentive to maximise belief in supernatural sanctions to encourage people to pay their tithes in full. Vermin usually disappear naturally and the church exploited this by delaying trials to make it appear that God dispensed justice. By optimising trial length it was possible to maximise agents’ posterior beliefs in God’s power. This explains why trials were so frequently delayed by the defence lawyers for vermin and why time series analysis suggests that these kangaroo courts were used more frequently at times of high rates of heresy as measured by the incidence of witch trials.

The witness getting gorilled

The witness getting gorilled


The US has not completed a new set of trade negotiations since 2007 when the KORUSFTA was agreed. Despite this Obama’s accomplishments in trade will likely be remembered as some of the most important achievements of his administration. Under his Presidency he secured Congressional approval of the Korean, Panamanian and Colombian FTAs. However these pale in comparison to the progress that has been made in negotiating the Trans-Pacific Partnership (TPP). With the stagnation of the Doha Round, future progress in trade liberalisation depends on such agreements. Unfortunately US Congressional recalcitrance may destroy American involvement in the pact.

The Peterson Institute for International Economics estimates that the TPP track of negotiations could lead to gains to US GDP of 0.4%, with one third of these gains attributable to investment provisions. NZ would gain some 2.3% of GDP. However the largest winner would be Vietnam who would be 13.6% better off. TPP negotiations have also prompted China to begin trade negotiations with its neighbours, a process that could lead to an eventual regional FTA that would boost US output by 1.5% and Vietnam’s by 21.5%.  In the medium-term the TPP would act as an important extension of Obama’s foreign policy pivot as argued in this excellent Op-Ed from Roger Altman and Richard Haass, the President of the CFR. Bizarrely although perhaps not unexpectedly Bernie Sanders cites the benefits that this will bring to Vietnam and similar countries as reasons to oppose the agreement:

The TPP will make the race to the bottom worse because it forces American workers to compete with desperate workers in Vietnam where the minimum wage is just 56 cents an hour

This comes in addition to an odd claim that Vietnam will export contaminated fish to the US. It’s not clear why Sanders arrives at the position that our primary moral obligations lie with the bottom 99% of the top few percent of the global income distribution. In order to have any chance of completing the TPP, Congressional approval of TPA (Trade Promotion Authority) is needed. This allows any negotiated agreement to be put to a simple vote by Congress with no amendments or filibustering. This is analogous to a bridge-burning strategy as it makes other parties willing to make better offers without being concerned of the risk that further talks may be needed as a result of some house amendment over an aspect of the deal that has already involved substantial compromise from all parties. Given the vocal opposition to the agreement from within the populist wing of the Democratic Party it is particularly important that TPA is approved if there is to be a TPP at all. TPA has passed both the House and the Senate. However the Senate bill was passed with TAA (Trade Adjustment Assistance), which provides some benefits to workers made unemployed as a result of not being efficient enough to compete following the passage of an FTA.

Vox has an excellent explanation of the situation in Congress. TPA only passed the Senate due to Democrats being willing to vote for it with TAA. The situation in the House is more complex. Republicans tend to support TPA but oppose TAA for ideological reasons. Democrats tend to support TAA given TPA is passed but are willing to vote against TAA in order to ensure TPA fails and the broader TPP fails. The original hope was that Republican support would ensure TPA passed and Democratic support would ensure TAA passed, thus reconciling the bills in the House with the combined bill in the Senate. Considering it appears that TPA would only be passed with TAA, one possible option would be for Republicans to approve TAA in the House to guarantee TPA – however the number of Republican votes needed for this makes this option very unlikely. A more likely but still very difficult option relies upon exploiting the fact that House Democratic opposition to TAA is not a subgame perfect equilibrium where the Senate has first mover advantage. If Senate Democrats pass a bill with TPA but not TAA, TPA would be guaranteed. Given this it would now be optimal for House Democrats to support TAA, which would allow for a later bill that could pass both houses. The risk in this strategy should be apparent.

Hillary Clinton is in a difficult position, as the moderate candidate in a primary contest against competitors who have an incentive to paint her as a DINO. Clinton has been one of the main supporters of Obama’s trade agenda and has played a huge role in helping the TPP move forward as Secretary of State. Unfortunately she now has to play the Democratic primary game, as this interview shows:

Jon: Last question and hopefully it’s a simple yes or no, but I’m not that optimistic.  If you were in the Senate still, would you vote for TPA when it gets there?

Hillary:  At this point, probably not because it’s a process vote and I don’t want to say it’s the same as TPP.  Right now I’m focused on making sure we get trade adjustment assistance and I certainly would not vote for it unless I were absolutely confident we would get trade adjustment assistance.

Similar dynamics are present in the Republican race, with Marco Rubio being placed under pressure by Rand Paul. This particular attack concerns the lack of public access to the negotiating text, something necessary to ensure parties to negotiations are able to save face in backing down over desired provisions without losing broader credibility in other arenas. Oh, and apparently the TPP, supported more by Republicans than Democrats is now being dubbed ‘Obamatrade’ by Republican populists – fantastic.

Much opposition to the TPP comes from its likely inclusion of provisions for ISDS (Investor State Dispute Settlement), which allows for arbitration to protect investments in countries with variable standards in enforcing the rule of law. Oddly many of those in the UK who oppose the parallel TTIP due to the inclusion of ISDS are strong supporters of the EU. The US has never lost an ISDS case despite many petitions against it – this underlines the fact that ISDS is about providing assurance that commitments to foreign firms cannot be arbitrarily overruled. Left-wing critics may be surprised to learn that the TPP actually provides for many basic ILO standards that aren’t properly implemented in some Pacific Rim countries – these guarantees are meaningless without ISDS-type provisions. In 2012 the Argentinian government siezed 51% of the assets of a Spanish oil company’s Argentinian assets. ISDS helps to stop such blatant state theft. This does not stop the ability of states to legislate for labour, environmental and health standards but requires compensation in cases of discrimination against foreign companies and reneging on prior agreements. Another case (Pacific Rim v El Salvador) currently before the ICSID under the CAFTA agreement concerns an American mining company who had its mining licence revoked despite earlier negotiations with the government and meeting all relevant environmental regulations. The Director of the National Economic Council has a thorough response to Sen. Warren’s concerns about ISDS. In particular the take-down of the examples she uses is clinical and on-point

In the only U.S. case alluded to in Senator Warren’s op-ed, a regulation was adopted banning a chemical used in gasoline additives made only by a Canadian company. An arbitration panel found in favor of the government and underscored the right of governments to regulate for public purposes, including regulation that imposes burdens on foreign investors, and noted that investors cannot expect that environmental or health regulations will not change. …

The Swedish suit against Germany is also instructive. Here, too, details on the case are not public, unlike cases under U.S. agreements. But available information suggests that the case is not about whether Germany can change its national energy policy to do away with nuclear power, but whether Germany needs to provide compensation for abrogating its existing commitments. German domestic courts have upheld claims relating to these issues in cases filed by Germany’s domestic energy companies under German law.

The reality is that the 3000 agreements today that include ISDS provisions (including in every single TPP nation) have not lead to the spectre trade opponents claim. Instead they are a crucial part of modern trade agreements that ensure a basic set of rules that encourage FDI and ultimately increase living standards.

Helping Small Business

The UK Labour party have promised to cut rates for small businesses and fund this by increasing the overall corporation tax rate from 20% to 21%.

We will cut and then freeze businesses rates for 1.5 million small business properties with a rateable value of less than £50,000. Labour’s plan means the tax burden on small businesses will be lower than under the Tories. We will look to go even further as we prioritise small businesses for future tax cuts.

During election year all politicians seem to want to talk about helping small businesses. This focus seems odd but is one that is questioned all too infrequently. For some reason we imagine small business owners to be stoic individuals at the cutting edge of economic growth. This assumption is strange. Research by the Brookings Institute found that more than 50% of small business owners founded their businesses for non-pecuniary reasons eg “wanting to be their own boss”. In contrast just 34% cited pecuniary reasons as their main motivation. On the other hand it would be difficult to find a CEO of a FTSE 100 company who opted for that job to live an easy life.

According to the 2014 UK Business Statistics Report 0.1% of firms in the UK provide 40% of the employment and produce 53% of turnover. 76% of registered businesses in the UK have no employees and produce 7% of turnover. Many of these firms would be offered tax breaks under Labour’s plans. In addition it seems much more likely that small firms will opt for hiring ‘cousin Joe’ when compared to large firms with stronger shareholder accountability for HR practices.

Research from the Kansas City Fed shows that data about job creation in small business can be misleading. From 1990-2003 small businesses were responsible for 80% of net new jobs in the United States but only 30% of gross new jobs. This suggests the apparent success of small business can be traced to job losses in large businesses rather than any ability of SMEs to create jobs relative to their larger competitors. The Bureau of Labour Statistics found that in 2004 businesses with less than 100 workers paid almost 25% of workers less than $8/hr as compared with 3% of employees at businesses with more than 2500 workers. In addition wage growth is higher at large firms suggesting this gap will increase over time. These effects remain after adjustment for a range of factors including fringe benefits (which also tend to be lower at small firms), demography, firm type and monitoring levels. One possible explanation for part of this effect comes from relative levels of unionisation. Collective bargaining is more likely to be implemented and accessible to workers at large firms. If Labour’s intention is to undermine union power and allow wages to fall to more competitive levels encouraging a shift from large to small business might be an advisable strategy. I doubt this argument will be convincing to many core Labour supporters.

In 2011 small firms in the US paid their workers wages equal to an average of 66% of the levels paid in large firms, with this gap widening over time. This is despite small firms being less likely to hire traditionally low wage groups such as women and ethnic minorities. Women make up 48% of the employees at large firms compared to 45% of those at small firms. Despite the large corporate world often being criticised for gender representation, small businesses do not appear to be different in the composition of their leadership. In 2012 18% of SMEs in the UK were led by women, which is in line with the 22% of women on FTSE 100 company boards.

Obviously the large firms of tomorrow will include many firms that are small today. However we should not confuse making an environment that encourages and incentivises these firms from one that provides handouts to the bulk of firms who have no real prospect of reaching these levels. A larger chunk of total technological innovation probably comes from refinement and perfection than big picture disruptive ideas. While we need the latter large firms have a clear edge on the former. American firms with 5000 or more employees spend more than twice as much per employee on research and development when compared with those with 100-500 employees. It seems more likely that large firms producing refinements of existing ideas are more sensitive to tax changes than Stanford drop-outs pursuing their Silicon Valley dreams. With venture capital becoming an increasingly important source of funding for start-ups it is now more likely than ever that the large firms of tomorrow are able to be picked up from amongst the vast majority of small firms which are going nowhere.