Thursday, 27 April 2017

Contra Dube & Harish, Queens are less good at both belligerence and baby-making than Kings.

Recently there have been a number of click-bait articles- and now even a Daily Mail piece- based on a Junk Social Science drive by regression utilizing weak and invalid instruments, claiming that Queens are more belligerent than Kings.

The conventional view was that Kings made War while Queens made babies.

In this post I will show that, for Europe between 1480- 1914

1) Kings made more Royal babies than Queens
2) Kings were more belligerent than Queens.

In other words, women were inferior both at making babies and making War.

Louis XIV said 'A King can only make a baby of the Royal Blood with the Queen'. According to this view, Kings can't make more Royal babies than Queens.

However, before his death, he himself raised 2 of his bastards into the Royal line of Succession. Charles XIII of Sweden went one step further. Without aid from any woman, at the advanced age of 60 he become the proud father of, first, a fat Danish Prince who promptly had a stroke and died, and then a French soldier of about 50 years of age. All this was done without the assistance of any women whatsoever. This shows that Kings made more Royal Babies than Queens over the period in question.  By contrast, Queens were vaginally constrained in this matter and, speaking generally, weren't able to legitimise bastards absent a complaisant Royal husband.
Thus, the stereotype of females as being necessary for making babies is shown to be mere Feminist propaganda, at least as far as European Royalty was concerned.

Secondly, Kings were more belligerent than Queens because only Kings have created their own Kingdoms and, almost invariably, have done so by purely military means. Napoleon created a lot of Kingdoms for his family and comrades in arms and about 4 million people died as a result. No woman has achieved anything remotely similar.

Sometimes a daughter or wife of a King proved more able than any man and managed to preserve the dynasty and expand its territories. However, such instances were rare and heralded no great change in the status of women. In other cases, where a woman was the sole legitimate heir, she might- if she survived the hazards of serial child-birth- enjoy greater longevity than a man would have done and this might correlate to what appears to be a more intensive involvement in belligerence. However, in this case, it is longevity, not any attitudinal or gender specific trait, which generates the result by means of an impact on the building up of State capacity- in particular, the ability to maintain long term alliances and honour treaty obligations.

Acharya & Lee have a paper arguing that, in medieval Europe, Kingdoms which had plenty of male heirs did better than those where such heirs were lacking and this had persistent hysteresis effects down to our own day.  One way to ensure plentiful male heirs is to increase gender dimorphism such that Queens concentrate on making babies and Social Conventions evolve such that Royal kids and women are immune from massacre. If male siblings co-operate, an even better outcome is feasible. Primogeniture with Religious sanction can help. Ottoman Turkey was weakened by the custom of killing or imprisoning the brothers of the Sultan on the legal principle of 'maslaha' (Public interest) because of Religious abhorrence of 'fitna' (Disorder) . By contrast, in Christian Europe, it was more commonly the case that the younger brothers of the Monarch were more Royalist than the King- and this sometimes created problems if they acceded to the throne.

Kingship can be sacerdotal in origin, but more commonly it arises out of military success`. War, proverbially, was the sport of Kings. Princes habitually dressed in military costume and, in extremis, took command of armies and hazarded all upon the outcome of battle. Queens, on the other hand, more especially under conditions of monogamy and primogeniture, were baby making machines. Their political value was a function of fertility not military prowess.

Sociobiology explains that male conquerors can have hundreds of babies and literally millions of descendants. Women- even Queens- can manage little more than a dozen kids. The pay-off, in terms of reproductive success, from conquest is far greater for the one than the other. Thus belligerence is a Kingly, not Queenly trait. On the other hand, being less tempted by conquest or the chance to hazard an existing domain in the hope of securing one more valuable, Queens may be more inclined to play the 'bourgeois strategy' and, it may be, this fact explains why females inherit in the absence of male siblings in some places and at some times but not others.

In the case of Europe, between 1500 and 1870- after which no Christian Queen truly ruled rather than reigned- two idiosyncratic factors are salient. Firstly, that the Balance of Power could be disrupted by dynastic marriages- as late as 1870 the Chancelleries of Europe worried over which Prince was marrying the Spanish Infanta, though, with hindsight, the whole thing was a chimera.

Royal Marriages were strategically important because of a peculiarity of Western Christendom- viz. insistence on monogamy and difficulties related to divorce in the event of the lack of a male heir (in the Eastern Church, divorce was permitted). This peculiarity was tied up to the role of the Church (which claimed to be the bride of Christ and thus insisted on monogamy without divorce) in legitimising secular power. After the Church split up, dynastic questions still had salience because of the Augsburg rule 'Cuius regio, eius religio  linking the religion of the monarch to that of the State. Ultimately, Nations asserted their right to chose monarchs of their liking and to constrain the monarch's actions.  However, this was an imperfect and sometimes dysfunctional process- more particularly in multi-ethnic polities or traditionally despotic ones like Tzarist Russia. Thus the great conflagration of 1914 had all the appearances of a struggle between crowned cousins though what was really of salience was ethnic cleansing and class warfare.
Ultimately, it became clear that Europeans couldn't be trusted to run their own continent and, in 1945, peace was restored by 2 new Guardians- one American, who still endures, the other predominantly but, alas!, not Asiatic enough, which consequently has fallen by the wayside.

A question we might ask is why some European polities, from the Fifteenth Century onward, adhered to the Salic Law- where no female could inherit the throne- while others did not. Before we do so, it would be well to point out that the Church's tables of consanguinity prevented an easy way to get round the Salic Law- viz. marry the female heir to the agnatic successor. However, the Church could be flexible. Thus, for example, the Pope was prepared to allow Henry VIII's bastard to marry his half-sister, Bloody Mary, and come to the throne that way.  Later on, the first undisputed Queen regnant of Portugal, Maria the Mad, who came to the throne in 1777, was married to her Dad's younger brother. Bu then the Church's ability to secure a rent by being flexible in this matter had declined in the same manner as its power to release vassals from their oath of obedience had previously disappeared.

Western Europe was peculiar in that Church and Crown had separate origins, legal codes, and motivating interests. Nevertheless, for a mobile 'warrior class', there was an advantage in something approaching 'matrilocality', mediated by a sacerdotal cadre. Here, the wives and daughters remain on the demense and extract resources from it with the assistance of the Clergy.  A mutuality of interest developed whereby the Church, which inherited disproportionately from women, had an interest in legitimising non agnatic inheritance.

The development of 'Civil Society' meant that a third branch of Law gained salience. This was linked to Parliament exercising a fiscal function conditional upon an increasing legislative role.

The Hapsburg 'pragmatic sanction' presents an interesting case where the last male heir disinherited his nieces in favour of his possible daughters. It is noteworthy that this succeeded because 'National' parliaments found it in their interest to affirm this departure from the dead hand of Salic Law. This marked the loss of salience of the 'Universal' Church and the growing importance of 'Nations'- as defined by Parliaments. Similarly, in 1717, the French parliament reversed the Sun King's edict raising two of his bastards into the line of succession of the French Crown- which would have had the effect of extinguishing what later became the Orleanist claim. The argument then made stressed the right of the French to choose their own King- thus anticipating the language of 1789.

Economic theory explains why 'transferable utility' is necessary so as to permit effective collective decision making under circumstances where authority is plural. Money is 'transferable utility'. So are women dowered with land. Indeed, assuming kin-selective altruism, and son preference for high status men and daughter preference for lower status men who can buy a high status husband, the course of Western European Social History is in line with what Economic theory would predict.

What Economic theory does not predict, because it is impossible, is that Queens will be more belligerent than Kings.

Yet, 2 American Professors- Dube & Harish- make this foolish claim in a much cited paper. 
The abstract reads-
'Are states led by women less prone to conflict than states led by men? We answer this question by examining the effect of female rule on war among European polities over the 15th-20th centuries. We utilize gender of the first born and presence of a female sibling among previous monarchs as instruments for queenly rule. We find that polities led by queens were more likely to engage in war than polities led by kings. Moreover, the tendency of queens to engage as aggressors varied by marital status. Among unmarried monarchs, queens were more likely to be attacked than kings. Among married monarchs, queens were more likely to participate as attackers than kings, and, more likely to fight alongside allies. These results are consistent with an account in which marriages strengthened queenly reigns because married queens were more likely to secure alliances and enlist their spouses to help them rule. Married kings, in contrast, were less inclined to utilize a similar division of labor. These asymmetries, which reflected prevailing gender norms, ultimately enabled queens to pursue more aggressive war policies.'

How do Dube & Harish achieve this result? Well, they exclude those parts of Europe which had few 'natural frontiers' or barriers against invasion and featured overlapping territorial claims- viz. France, Germany, the Balkans and so on- because, since bellicosity was geopolitically required, they had strict laws against female rulers or inheritance through the females, though, in the case of Poland (which Dube & Harish list as a 'non Queen polity') there actually was a Queen Regnant from 1575 to 1586.

Instead, Dube & Harish concentrate on peripheral parts of the continent, the Iberian peninsula, Britain, Sweden, Naples etc. This means their work is ab ovo worthless because it is not comparing like with like.
Moreover, their list of 'non Queen polities' is historically inaccurate-


Bourbonnais was Salic- except for a brief period of less than 20 years when Suzanne, the daughter of Anne of France, ruled the Duchy. After her death, the mother of the French King offered to marry her widower so as to get her hands on the Duchy. He refused on the grounds of her advanced age. Her son promptly stripped him of his titles and possessions. So much for Bourbonnais being a 'Queen Polity'.
But then Dube & Harish aren't sticklers for the truth. If Bourbonnais, a Duchy, which had Duchesses not Queens can count in their map why not Polities which did actually have Queen regnants?
Poland most definitely had a Queen- Anna the last of the Jagiellons- but it is coloured as non-Queen polity.
Maria Theresa became 'King' of Hungary and Croatia, and, by conquest, Queen of Bohemia. But still all these vast lands are coloured as 'non Queen polities'. She made her husband a co-ruler of these 2 Kingdoms and by a combination of force and bribery had him recognised as Holy Roman Emperor.

Another Austrian princess or superior ability was Maria Carolina of Naples. Once she gave birth to an heir, she was able to get the better of her father-in-law, the King of Spain, and promulgated a relatively progressive reign over which her dunce of a husband nominally presided.

Thus, the truth is, Queens have ruled the roost even in 'Salic' lands. France had Female Regents, like Anne of France- who was able to make Suzanne a Duchess in her own right- Catherine de Medici and Anne of Austria- that last notably less prone to aggression than either her husband or son.

A regnant Queen might not have any actual power if she were mad- like Maria the Mad of Portugal, or was under the thumb of her husband, like England's Mary II, or simple-minded and in thrall to a charismatic courtier, like Queen Anne, or a mere puppet for her maternal Uncles, like Lady Jane Grey, or else because the Crown itself conferred little tangible power, like Queen Anna of Poland, or Queen Victoria of England,  or because a religious or moral scruple militated against it and so some powerful prelate or confessor pulled all the actual strings.
Equally, a Queen by marriage might actually exercise all the functions of a regnant Queen either officially- as Regent, for an absent or incapacitated husband or minor son, nephew, or other heir- or else unofficially by reason of an uxorious husband content to leave power in the hands of a 'petticoat government'.
One unusual case is Empress Catherine of Russia- a servant girl of chequered past whom Peter the Great elevated to co-ruler status. After his death, she was able to rule on her own for a couple of years, with the support of the 'new men', before dying of natural causes. Her daughter, Elizabeth was able to usurp the throne some years later and, as such. can be considered the only woman to have gained the Crown entirely by her own initiative. The case of Catherine the Great is more complicated. She hated her husband but, as a foreigner, probably could not have pulled off a coup by herself. It is a matter of conjecture as to whether these Tzarinas were more or less bellicose than would have been the case if a Tzar had held the throne. Catherine's son and heir certainly wrote critically of an expansionist policy and initially suited his actions to his words after his succession; her husband too might plausibly be presented as a 'peacenik'. On the other hand, Peter III, like Paul I, had a fascination with the drill-ground and his odd views, extra-territorial interests, and propensity to hero-worship were such as were likely to involve expensive and quixotic campaigns abroad. Few doubt, on the other hand that Catherine had superior administrative skills, a more balanced character and sufficient diplomatic finesse to be an effectual international mediator.

Had Dube & Harish been serious about wishing to measure the effect of a Queen running things on the incidence of warfare they might start by picking the set of Queens (Q) considered by Historians to have had actually held power and then looking to see if there was any increase or decrease in incidence of Warfare (W) against the trend rate. The problem here is that we might want to distinguish between the effect of a King not wielding power and a Queen doing so. Suppose, in the absence or incapacity of a King, the norm is for a Regency Council to run things and that Queens in this case exercise no power even if they are regnant, then, it may be, there is one effect on W (War incidence). We need one instrumental variable to capture this effect which is about the increase or decrease in war resulting from not having a King running the show. If males, on average, turn into testosterone fuelled monsters once they get their tushy on a Throne, the 'no King' effect would yield a negative correlation with W.

Now suppose there is no Regency Council but, instead, a Queen runs things. We need a second, completely independent, instrumental variable to capture the effect that her gender has on W. Suppose warriors are prejudiced against females. They think a Queen running things is bound to be crap at defending her territory. Then, the 'Queen effect' would yield a positive correlation with W relating to more defensive wars.

 Taken together, these two instruments would ideally pick out every casus belli in W which directly relates to the Queen's gender. Since not all Queens or Kings are alike, what we are looking for is the average effect of having a Queen run things. The problem here is that if people have rational expectations, or, at the least, they learn from experience, then, very quickly- if we would find after running the best possible Econometric investigation that there is a zero Queen effect on W- there will in fact be a zero Queen effect on W because everybody will anticipate this outcome. However, because expectations are so important, the Q series couldn't have been independent of the W series and vice versa. What is happening here is rubbish Econometrics.

To their credit, Dube & Harish aren't even trying not to do rubbish Econometrics as opposed to click-bait Junk Social Science.

They choose the following 2 instruments
1) 'we utilize whether the first-born legitimate child of the previous monarch(s) was male as one of our instruments for whether a queen holds power.'
Consulting the table of instruments used by Dube & Harish we come in for a surprise. Empress Elizabeth of Russia is excluded from the list of Monarchs, though she usurped the throne, but is included as an instrument for Catherine the Great! Why? There was no blood relationship between them. Catherine threw her hubby in prison and took the throne. Why is she shown, by this instrument, to come to the throne as if she were a niece of Elizabeth?
It appears Dube & Harish inhabit an alternative Reality with its own History
Mary II of England is unusual in that her father was alive when her husband usurped the throne in their joint names. The reason he did so was because her father had just had a son. In other words, this is a case where a woman becomes Queen because her father has a son.
Pretty complicated right? How do Dube & Harish deal with her case? This is the table they give in the appendix-
I'm baffled. Where is Mary II? We can see her husband inheriting sole rule at her death. But what about her? Why treat her case as if it was routine? Moreover, by the Bill of Rights, the throne would have gone, not to Anne but to Anne's son had either William or Mary lived into his majority. In the event he died before William.  None of this explains why Mary features as 'instrument ruler 1 for her sister Anne. The truth is, had the 'Old Pretender' converted to the Protestant faith, he would have inherited from his aunt.
Dube & Harish supply this explanation
This makes things as clear as mud. I suspect there is some fraud at the bottom of it.

Consider the following snippet from their table of instruments regarding the Duchy of Lorraine-
Henry II tried to give the Duchy to his daughter who married the son of a guy who pressed his superior Salic right and got the Dukedom before consigning it to his son who hated poor Nicola and abandoned her and married again in defiance of the Church. He was chased out of his Duchy which he gave to his younger brother who gave it back after being chased out by the French within a few months. Charles V was the son of this brother. He was Duke in name only.   However, in Dube & Harish's fantasy world, Nicola is an instrument for his succession to a non-existent throne.  Charles IV did get restored for a few years  before being chased out again. By the time of his death, he had legitimated his own children. They would have got the Dukedom if it was worth getting. It wasn't. Charles V's loyal military service to the Austrians did however get his son the Duchy later on. This son had a son who gave up Lorraine to marry Maria Theresa.

I don't know if this playing fast and loose with instrumental variables is the result of fraud or just ignorance on the part of the authors. In either case, it invalidates their result.

A good instrument is one which correlates well with the explanatory variable (in this case, Q) but which is exogenous, i.e. independent, in respect to the 'error' term (i.e. explains nothing with respect to W).
Dube & Harish's instrument does not correlate with Q at all. It is far too weak. What matters under male preference (semi Salic) primogeniture is that no surviving male sibling exist for a Female to gain the throne in her own right. Gender of the first born is irrelevant.

In any case, 'First born legitimate child' is not well defined because Kings had the option to decide that an illegitimate child, regardless of birth order, was in fact legitimate and a legitimate child was in fact illegitimate. Henry VIII decided at one time that Henry Fitzroy was legitimate and that Mary was not because his marriage to her mother was invalid.
However this decision of Henry's had a hysteresis effect on W. So there is at least one case which shows that, methodologically speaking, this instrument must be discarded. Moreover, Kings who were not the 'first born legitimate male heir' have often fared badly. England's Charles I, James II; France's Charles X- there are numerous instances where the younger son is more Royalist than the King and his stiff necked ways bring down the dynasty.

Dube & Harish, however, are innocent of any notion of European history. More remarkably, they are also unfamiliar with any human society which has ever existed. They write-  'The lack of a first born male could spur war if it signals uncertainty in succession. Other monarchs may choose to attack the polity if they see that the first birth did not yield a male heir. If so, queens would inherit polities that are already participating in more wars, which would present an alternative path through which the instrument affects war participation.'  Wow! A King is either weak enough to be attacked now or he isn't. Typically, a King gets married in his twenties and starts having babies. If his first born isn't male, the fact will be known by his Thirties. Will anyone in their right mind attack a King when he is at his peak as a military commander? Suppose someone is foolish enough to moot the idea. He will get no support. Why? Because the fact that the first born is a girl does not mean all subsequent kids will be girls. What if the King has one daughter and is known not to be able have any more children? Does this create a problem? No. If the Kingdom follows Salic law, the Heir presumptive is easily identifiable. Sometimes this Heir will be married off to the daughter regardless of the tables of consanguinity. Alternatively, if the Kingdom is semi-Salic, the daughter can be married to a Prince of proven martial or diplomatic worth. In no case does the fact that a man's first child is female raise any suspicion in anybody's minds that he will never father a son.

 Dube & Harish may not know this but ordinary people do. By contrast, infertility of the monarch- by reason of constitution or lifestyle- could lead to predatory behaviour by foreign powers. Schleswig-Holstein followed Salic Law whereas Denmark was 'semi-Salic'.  When it became clear that the King could not father an heir, the Danes chose a Crown Prince who had a plausible claim to the Duchies so as to preserve the Personal Union. Previously, the legitimate heir to the Duchies had been bought off and had transferred his claim to the Danish Crown. However, his son revived it and, propelled by German Nationalism, and Prussia's desire to build the Kiel Canal through Duchy of Holstein territory, a War was fought on this issue.

This is a clear case where the childlessness of a King resulted in a predatory war. Thus, the infertility of a Monarch can be correlated with W. On Dube & Harish's account, 'First born is female, which we have seen is not correlated with Q, ' is in fact correlated with W. But, in that case, it is an invalid instrument because it is endogenous.

What about the notion that a country will be attacked the moment a Queen is put on the throne? Since War is costly, nobody in their right mind would put a Queen on the throne under these circumstances. If the stakeholders in the polity are not in their right mind, or if they hope to personally profit from a general catastrophe, then the polity will succumb to Darwinian forces. It will be eliminated. In this case, it was some problem in the polity, not the gender of the monarch, which is the causative factor.

Where few natural or ethno-cultural barriers to invasion exist, something like a Salian law or an elective Monarchy would be adaptive for the polity and we do find that Salic parts of Europe had this feature.

Dube & Harish aren't testing for what they say they are testing for. They are testing the proposition 'European polities between 1480-1918 lacked rational expectations and made crazy assumptions like thinking if a King's first baby was a girl this meant all his kids would be girls and so he should be attacked immediately.' Naturally they find that the proposition is false because it is one only they were stupid enough to suggest.

2) The other instrument used by Dube & Harish has to do whether the previous monarch had a sister. Typically, they show Lady Jane Grey as Elizabeth's sister though there was no relationship between them.  Elizabeth was Mary's sister but Mary has been put in as the first instrument as though she were actually her father. God alone knows if this is stupidity or fraud.

Consider the following explanation- 'Of course if the previous monarchs did not have any children, or the children died by the time of accession, or were too young to rule at the time of accession, then the throne could pass to a sibling of the monarch instead. If the previous monarchs had a sister then the throne could pass to her, as she would be given priority over more distantly related males. For example, Ulrika Eleanora became ruler of Sweden in 1718. She was preceded by her brother Charles XII, who never married or had children. In addition, all of their brothers had died by the time Charles’ reign drew to an end, leaving Ulrika as the heir. Since having a sister enhanced the chance of having a female accede, we also use whether the previous monarchs had a sister as a second instrument for having a queen in power.'
Dube & Harish are telling porkies. Ulrika was not the heir. Her elder sister's son,  Charles Frederick of Holstein-Gottorp, was the heir. Ulrika curried favour with the Swedish Parliament, with whom she had enjoyed good relations as Regent, by ending absolute monarchy and making the Crown elective. The reason she cheated her nephew out of his rights was because she was madly in love with her husband in whose favour she abdicated within a couple of years. She did not rule anything. She was a girl in love. Her brother- the famous Charles XII- loved fighting. She did not. Her husband was tolerated because he was weak and uninterested in politics. Sweden ceased to be a 'big power' in military terms thereafter. Her husband, an adulterer, had no power, though he transferred resources from his own patrimony to his new 'Kingdom.  Later the Swedes blundered into a War with Russia which they lost. Empress Elizabeth insisted that Ulrika's nephew's line be reinstated as her husband's successors- the poor chump had impoverished his patrimony in vain.

So much for Dube & Harish's second instrument. It clearly does not predict that a Queen will hold power because in the one instance they supply, the opposite happened. Ulrika lost power. The Crown lost power. There is no 'regnant Queen' here at all.

This is not Dube & Harish's view. They say 'However, the OLS estimates in Table 3 may be downward biased — for example, if the elite allowed queens to come to power more during times of stability, or prevented them from coming to power during times of war. In fact, even some reigning queens articulated the view that women should not govern if they had to lead armies into battle. This was the position of Ulrika Eleanora who asked that the Swedish Riksdag that her husband Frederick be made co-regent'. As we have seen, Ulrika was appointed Regent by her brother, an absolute monarch, who was away fighting. The Riksdag had no power to appoint anyone a co-Regent. After her brother's death, Ulrika claimed the Crown, to which she was elected after she ended Absolute Monarchy and inaugurated the 'Age of Liberty'. Since Swedish Law did not permit two people to rule, she abdicated in favour of her husband not because she was so stupid and ignorant as to think a Ruler had to 'lead armies into battle' but because she was in love.

Dube & Harish aren't historians. But they do have access to Wikipedia. This paper of theirs has been circulating since 2015. Why are they so careless and stupid? There is scarcely a single historical example they cite which does not rebound against their own argument.

They don't understand that Western Europe started off with 'limited monarchies' but moved in an absolutist direction before being pulled back again by Parliaments.

By the  middle of the nineteenth century, Queens generally reigned but did not rule.

They write- ' Prince Albert was Queen Victoria's most trusted advisor, and shaped her colonial policy and public relations image (Urbach 2014). In fact, Victoria was said to be most active as a ruler during Albert's lifetime.' Queen Victoria had no political power. Her Uncle was the last British monarch to appoint a Prime Minister against the will of Parliament. Victoria had no 'Colonial Policy'. Even Disraeli, who liked flattering her coz he was a great big Queen himself, never dreamt of accepting any suggestion of hers- e.g. permitting Indian Princes to enter the House of Lords.

Dube & Harish also mention Queen Dona Maria II of Portugal, a contemporary of Victoria, who was put in by the Liberal faction for the specific purpose of carrying forward a constitutional monarchy of the English type. Her husband, the Catholic son of an Uncle of Victoria's, was equally committed to Liberalism and later refused the Throne of Spain.  Neither was belligerent or autocratic. The dynasty itself ended in 1910 because it didn't really make much difference to anybody. By contrast, Maria the Mad is remembered favourably in Portugal and Brazil despite her horrible shrieking. Like Mad King George, she pursued true 'Raj Dharma'- the true duty of a King- viz. to bring Enlightenment to the Masses the only way Kings, or Professors of Economics, can do so- viz. by screaming incessantly and shitting themselves at all opportune moments.

Dube & Harish believe there is a 'division of labour' such that Queens get their hubbies to do useful work, whereas Kings just keep Queens barefoot and pregnant. Thus, Queens have more resources for war. This is idiotic. States simply do not face the sort of constraint that a middle class household does. If my wife forces me to take a job, sure, my family will be better off. If the Queen insists Prince Phillip get a job, Britain's prospects in War or Peace are not affected at all.

Leaving aside the limited Monarchies of Western Europe, there is only one Christian Monarchy which greatly expanded its domain under female leadership. This was Russia where, uniquely, an unmarried woman seized power and ruled in her own name- not as the regent of an existing or possible son. This was  Empress Elizabeth, who was followed by Catherine the Great, whose hubby was off his chump and no use at all to anybody. Neither fits in with Dube's & Harish's scheme. Indeed, even Empress Anna- who created the precedent of female rule in Russia, came to the throne contrary to the rule of succession, or Dube & Harish's instruments, because she promised to give more power to the Privy Council and was unencumbered with a husband. However, she soon reneged and re-established Absolutism, forcing Prince Golitsyn to marry a Kalmuck maiden in an ice palace while being taunted by dwarfs and zanies.

Another thing Dube & Harish ignore is whether or not a Queen has a son, for whom she can act as regent. Isabella of Castile is an example of a Queen who only became legitimate and could carry forward her policy after she had a son. She had already demonstrated bravery and capacity to lead, but that's not what clinched her prevailing over other possible female heirs.

Dube & Harish write-
Why is this first stage necessary?
The Historical Chronicles clearly state when and for how long Queens held power.
If the fact that a Queen is running the show is correlated with military challenges to the polity, any 'informative' instrument (i.e. one which correctly predicts a Queen will rule) will not be 'valid' (i.e. uncorrelated with military challenges). However, 'conditioning'  can create a Statistical artefact- i.e. can mislead. So there is no point, except to wilfully mislead, in putting in such an instrument.

Dube & Harish's equation is a piece of shit.

Take the case of Elizabeth of Russia-  the only woman to have seized the throne and  to have done so in her own right, without a husband or son. The previous rulers had been her more or less distant cousins.  The equation given by Dube & Harish features an utterly useless 'instrument' in her case. 
True, she was the daughter of Peter the Great, but there had been several Monarchs in between. Elizabeth was a fine diplomat and may have turned out to be a bellicose leader but she fell ill just at the moment when this possibility could fructify. She is unique in that she was never married to any Prince and had no children. She was a truly self-made woman. However, her passion was for extravagant balls, not battles.

 An instrumental variable is only used when the explanatory variable (in this case, whether or not the Ruler was a woman) can itself have an effect on the dependent variable (war-related outcomes in the polity). But the instrument should correlate with the explanatory variable. We can see that for Queen Regents or usurpers like Elizabeth of Russia, Dube & Harish's instrument does not correlate at all. Why use it?

Now it becomes clear.
Dube & Harish are not just being stupid they are betting Professors who read it will be equally stupid.
Why?
Professors, at least of Econ, in America are middle class.
They don't understand anything about the European Aristocracy.
There are Zero cases in the period under consideration when ' war in past reigns led children to die young' because we are dealing with Royalty- not peasants. During wars, ordinary people have to do extra jobs to make ends meet. Still their kids may die at a higher rate. This does not happen to Royals. A Queen does not need to get her hubby to take a job so the country will be better managed. A King does not lose a lot of his little kiddies coz they starve to death or get raped or bayoneted by enemy soldiers.

What Dube & Harish have written is nonsense on stilts but it looks like 'Econometrics'.
A far better approach would be to choose the following two instruments- 
1) The Ruler being widely believed to have a vagina
2) The Ruler being widely believed to lack testicles.
Such 'instruments' correlate with being a Regnant Queen and are exogenous for Statistical purposes since large scale wars are not fought because a Vagina, as opposed to pair of testicles,  reposes upon a particular Throne's cushion.  Dube & Harish's instruments do not meet this criteria because, unless polities are cohesive, there can always be conflict about who was the true first born son, for the purpose of legitimising descent under primogeniture, or deciding which sister should rule.

Their data is rubbish because we don't know the gender of any Royal's first born because it was unlikely to be legitimate- though, it could be rendered so at a later date. In the 15th Century, John the First of Portugal and Ferdinand I of Naples were bastards.  The status of Elizabeth I of England was equivocal. The Duke of Monmouth was at one time expected to be legitimised and, had he not rebelled, he might well have become King after his Uncle James II had exhausted the patience of the Commons.

It is crazy to worry about finding out the gender of who was or wasn't the first born simply so as to have an 'instrument' which looks 'exogenous' because the said instrument is in no way correlated with whether or not a Queen was running things. 
What Dube & Harish are doing here isn't Econometrics, it is play-acting of a particularly stupid and ignorant sort.


These 'controls' are worthless.
What matters is that a polity is cohesive enough to decide issues of legitimate succession itself.
If it does so, it does not matter if a Vagina or a Penis is cushioned by the Throne.
If it doesn't, what matters is the luck & cunning of the incumbent which in turn does not depend on whether they possess a Vagina or Penis.

We know this because, in recent years, Europe has been turning to pure primogeniture precisely because polities are cohesive and War unthinkable. This in turn impacts on fertility. We don't expect to see a preference for a male 'heir & spare' to continue and this means lower expected Royal family size.
In other words, what Dube & Harish are actually testing for has nothing to do with whether Queens are more or less likely to be belligerent. They are testing to see whether they can predict Queenly rule. They can't. Nobody can. But we don't need to. History tells us whether or not a particular country was ruled by a Queen or not. What determined whether a particular Queen was belligerent also determined whether she became Queen in the first place. This has to do with expectations w.r.t her adherents prevailing. Just on this basis, we would expect to see some increased belligerence, ceteris paribus, from regnant Queens. However, we would also expect to see the Queen's adherents taking steps to entrench their own privileges and thus creating or reinforcing 'limited monarchy' which by itself would damp down subsequent monarch dependent belligerence effects. In any case, War- as Clausewitz said- is only the extension of politics by other means. Selecting a King with a good military reputation may be enough, in itself, to secure all the benefits of War without its costs. Thus a humbly born French soldier was chosen to be King of Sweden purely on the basis of his military prowess. He persuaded Norway to continue a personal union under his Crown with Sweden. The House of Bernadotte still rules the latter country. In the event, the founder of the dynasty did not have to fight any battles. His reputation was sufficient.

By contrast with Sweden, the Brits, under long lived Queens like Victoria and Elizabeth II, both of whom reigned but did not rule, ended up invading or otherwise conducting military operations across countries occupying 90 per cent of the Earth's habitable surface. So what? Nothing can be deduced from this about Queenly belligerence just as nothing can be deduced from Mad Queen Maria's sojourn in Brazil. She didn't really endow that vast territory with Liberty. She shrieked loudly and soiled herself.

Dube & Harish write 'Using the first born male and sister instruments, we find that polities ruled by queens were 27% more likely to participate in inter-state conflicts, compared to polities ruled by kings.' They know that their instruments don't pick out Queens who exercised power as Regents, as can be seen from this footnote- 'France did have queen consorts who married reigning kings or queen regents who were essentially acting monarchs on behalf of child heirs who were too young to rule. Note that identifying the effect of queen regents would require a different empirical strategy than the one we use in this paper since gender of the first-born child and gender of the siblings of previous monarchs do not have predictive power in determining whether queen regents came to power.' Moreover, these instruments don't predict Queens like Catherine the Great while including Queens like Victoria who 'reigned but did not rule'.
Technically, Dube & Harish's instruments are not 'weak'- The bottom of Table 4 shows that the instruments together make for a strong first stage: the Kleibergen-Paap F-statistic is 13.7. Individually, each instrument also has a statistically significant effect on the likelihood of a queen coming to power. If the previous monarchs had a first-born male, this reduced the likelihood of a queen coming to power by 21.2%. In contrast, if they had a sister, this increased the likelihood of queen coming to power by 18.9%.- however, they could have got an even better result by choosing 'Ruler had a Vagina' and 'Ruler lacked Testicles'. We already know that most Queens inherited their position and so Dube & Harish's instruments are going to look kosher- but only to stupid Americans who think that European Queens got their Crowns by winning a pageant or because a Fairy Godmother waved her magic wand. The problem is that some other Queens, like Catherine the Great, did not inherit their position and actually ruled, unlike Mary II or Victoria who merely reigned. So what we have here is junk Social Science at its best- 'These estimates are economically important, representing a doubling over mean war participation over this period. In contrast, we find that queens were no more likely to experience civil wars or other types of internal instability.' 

Dube & Harish well illustrate the stupidity and ignorance of the Academy. Being of Indian origin, they have gone the extra mile to display racism as well as misogyny- thus, they cheerfully endorse Philip Quincy Wright's distinction between Civilised Europe and barbarians of darker hue by quoting his classification of Wars as follows- 
'

So, kids, what have we learnt today?
Women only get to Rule if their brothers die or were never born.
If married, they get their hubbies to help them coz they are shite in their own right.

What's more, the  'modern family of nations' is White and Christian.
Alien cultures aren't part of this family but Imperial Wars 'expand modern civilisation', which is a good thing coz dark people or Muslims or Hindooos or Confucians are uncivilised and not modern. 

Anyroad, that's what this pair of Indian origin American Academics want us to believe they believe.

Trump would be so proud.

No comments: