|
Read the rules in the OP before posting, please.In order to ensure that this thread continues to meet TL standards and follows the proper guidelines, we will be enforcing the rules in the OP more strictly. Be sure to give them a re-read to refresh your memory! The vast majority of you are contributing in a healthy way, keep it up! NOTE: When providing a source, explain why you feel it is relevant and what purpose it adds to the discussion if it's not obvious. Also take note that unsubstantiated tweets/posts meant only to rekindle old arguments can result in a mod action. |
On February 19 2013 08:32 Souma wrote:Show nested quote +On February 19 2013 08:25 aksfjh wrote:On February 19 2013 07:04 {CC}StealthBlue wrote:After debuting with great fanfare, an RNC-backed scheme to rig the electoral college in blue states to favor Republican presidential candidates is looking mostly dead. But, to quote the Princess Bride, “there’s a big difference between mostly dead and all dead.” Enter Pennsylvania, where an electoral vote bill is more than slightly alive.
State Senate president Dominic Pileggi (R) is pushing a plan to change Pennsylvania’s current winner-take-all allocation of the state’s 20 electoral votes to a system that apportions them proportionally by each candidate’s share of the statewide vote. Had it gone into effect before the 2012 election, President Obama would have won just 12 electoral votes to Mitt Romney’s 8. According to a memo Pileggi wrote debuting his new plan, the revamped system “much more accurately reflects the will of the voters in our state.”
Pileggi’s bill is slightly less far-reaching a power grab than bills proposed in states like Virginia that would have divided electoral votes by congressional district, which thanks to gerrymandering strongly favors Republicans even in states Obama won handily. This is by design: Pileggi pitched a congressional district split along the lines of the Virginia bill in 2011 only to face near-unanimous opposition from Republican members of Congress in the state who feared Democrats would pour millions of dollars into winning their individual districts. While Pileggi has shifted on the issue, a pair of Pennsylvania state representatives have also introduced a congressional split version since the election.
But the biggest difference that sets Pennsylvania apart is that each of the top leaders necessary to pass a bill have expressed support for changing its electoral college. Gov. Tom Corbett (R) strongly backed Pileggi’s 2011 effort. So did state House Majority Leader Mike Turzai (R), best known nationally for bragging that voter ID restrictions opposed by civil rights groups would help Romney win the state in 2012. By contrast, similar electoral vote bills elsewhere hit a wall this year after Republican statehouse leaders, key lawmakers, or governors indicated their opposition. Source I'm actually ok with this. I don't like the people that support it and why they support it, but splitting the EC based on popular vote makes sense these days. It does make sense, but only if all the states do it imo, and if they aren't under the effect of partisan gerrymandering. Imagine if only blue states did this or vice-versa, the other party would never win an election.
I only agree with it if it is done by a statewide or nationwide popular vote, and not by districts. States can't really be gerrymandered. Also, it's one of those things that both sides would be pressured to do once a solid number went that way. Finally, we don't even know how it would affect an election since spending in states would be drastically altered as some ceased to be "swing states."
|
2nd Worst City in CA8938 Posts
On February 19 2013 11:01 aksfjh wrote:Show nested quote +On February 19 2013 08:32 Souma wrote:On February 19 2013 08:25 aksfjh wrote:On February 19 2013 07:04 {CC}StealthBlue wrote:After debuting with great fanfare, an RNC-backed scheme to rig the electoral college in blue states to favor Republican presidential candidates is looking mostly dead. But, to quote the Princess Bride, “there’s a big difference between mostly dead and all dead.” Enter Pennsylvania, where an electoral vote bill is more than slightly alive.
State Senate president Dominic Pileggi (R) is pushing a plan to change Pennsylvania’s current winner-take-all allocation of the state’s 20 electoral votes to a system that apportions them proportionally by each candidate’s share of the statewide vote. Had it gone into effect before the 2012 election, President Obama would have won just 12 electoral votes to Mitt Romney’s 8. According to a memo Pileggi wrote debuting his new plan, the revamped system “much more accurately reflects the will of the voters in our state.”
Pileggi’s bill is slightly less far-reaching a power grab than bills proposed in states like Virginia that would have divided electoral votes by congressional district, which thanks to gerrymandering strongly favors Republicans even in states Obama won handily. This is by design: Pileggi pitched a congressional district split along the lines of the Virginia bill in 2011 only to face near-unanimous opposition from Republican members of Congress in the state who feared Democrats would pour millions of dollars into winning their individual districts. While Pileggi has shifted on the issue, a pair of Pennsylvania state representatives have also introduced a congressional split version since the election.
But the biggest difference that sets Pennsylvania apart is that each of the top leaders necessary to pass a bill have expressed support for changing its electoral college. Gov. Tom Corbett (R) strongly backed Pileggi’s 2011 effort. So did state House Majority Leader Mike Turzai (R), best known nationally for bragging that voter ID restrictions opposed by civil rights groups would help Romney win the state in 2012. By contrast, similar electoral vote bills elsewhere hit a wall this year after Republican statehouse leaders, key lawmakers, or governors indicated their opposition. Source I'm actually ok with this. I don't like the people that support it and why they support it, but splitting the EC based on popular vote makes sense these days. It does make sense, but only if all the states do it imo, and if they aren't under the effect of partisan gerrymandering. Imagine if only blue states did this or vice-versa, the other party would never win an election. I only agree with it if it is done by a statewide or nationwide popular vote, and not by districts. States can't really be gerrymandered. Also, it's one of those things that both sides would be pressured to do once a solid number went that way. Finally, we don't even know how it would affect an election since spending in states would be drastically altered as some ceased to be "swing states."
I'm just saying, if we're really considering this, it would be nice if a lot of states or the entire country adopted it at once, as not to skew results one way or the other. If California decided to do it and only California, suddenly Democrats would no longer be guaranteed all of those 53 electoral college votes, so it gives them a bit of a disadvantage. Likewise, if Texas were to implement it, then Republicans would lose a lot of their guaranteed electoral college votes, so I think if we were to implement this, it should be widely encompassing.
|
On February 19 2013 12:16 Souma wrote:Show nested quote +On February 19 2013 11:01 aksfjh wrote:On February 19 2013 08:32 Souma wrote:On February 19 2013 08:25 aksfjh wrote:On February 19 2013 07:04 {CC}StealthBlue wrote:After debuting with great fanfare, an RNC-backed scheme to rig the electoral college in blue states to favor Republican presidential candidates is looking mostly dead. But, to quote the Princess Bride, “there’s a big difference between mostly dead and all dead.” Enter Pennsylvania, where an electoral vote bill is more than slightly alive.
State Senate president Dominic Pileggi (R) is pushing a plan to change Pennsylvania’s current winner-take-all allocation of the state’s 20 electoral votes to a system that apportions them proportionally by each candidate’s share of the statewide vote. Had it gone into effect before the 2012 election, President Obama would have won just 12 electoral votes to Mitt Romney’s 8. According to a memo Pileggi wrote debuting his new plan, the revamped system “much more accurately reflects the will of the voters in our state.”
Pileggi’s bill is slightly less far-reaching a power grab than bills proposed in states like Virginia that would have divided electoral votes by congressional district, which thanks to gerrymandering strongly favors Republicans even in states Obama won handily. This is by design: Pileggi pitched a congressional district split along the lines of the Virginia bill in 2011 only to face near-unanimous opposition from Republican members of Congress in the state who feared Democrats would pour millions of dollars into winning their individual districts. While Pileggi has shifted on the issue, a pair of Pennsylvania state representatives have also introduced a congressional split version since the election.
But the biggest difference that sets Pennsylvania apart is that each of the top leaders necessary to pass a bill have expressed support for changing its electoral college. Gov. Tom Corbett (R) strongly backed Pileggi’s 2011 effort. So did state House Majority Leader Mike Turzai (R), best known nationally for bragging that voter ID restrictions opposed by civil rights groups would help Romney win the state in 2012. By contrast, similar electoral vote bills elsewhere hit a wall this year after Republican statehouse leaders, key lawmakers, or governors indicated their opposition. Source I'm actually ok with this. I don't like the people that support it and why they support it, but splitting the EC based on popular vote makes sense these days. It does make sense, but only if all the states do it imo, and if they aren't under the effect of partisan gerrymandering. Imagine if only blue states did this or vice-versa, the other party would never win an election. I only agree with it if it is done by a statewide or nationwide popular vote, and not by districts. States can't really be gerrymandered. Also, it's one of those things that both sides would be pressured to do once a solid number went that way. Finally, we don't even know how it would affect an election since spending in states would be drastically altered as some ceased to be "swing states." I'm just saying, if we're really considering this, it would be nice if a lot of states or the entire country adopted it at once, as not to skew results one way or the other. If California decided to do it and only California, suddenly Democrats would no longer be guaranteed all of those 53 electoral college votes, so it gives them a bit of a disadvantage. Likewise, if Texas were to implement it, then Republicans would lose a lot of their guaranteed electoral college votes, so I think if we were to implement this, it should be widely encompassing.
That would be ideal, but nearly impossible from a practical point of view. The best I would hope for is that 3 or 4 smaller states would do it every 1-2 election cycles, until the bigger states (like Texas and California) wouldn't feel they would cost somebody an election. Even in the case of those big states, however, they are all strongly leaning one way or another. At most, we're talking about a loss of 30%, so about 16 votes from California and 12 from Texas.
|
On February 19 2013 13:09 aksfjh wrote:Show nested quote +On February 19 2013 12:16 Souma wrote:On February 19 2013 11:01 aksfjh wrote:On February 19 2013 08:32 Souma wrote:On February 19 2013 08:25 aksfjh wrote:On February 19 2013 07:04 {CC}StealthBlue wrote:After debuting with great fanfare, an RNC-backed scheme to rig the electoral college in blue states to favor Republican presidential candidates is looking mostly dead. But, to quote the Princess Bride, “there’s a big difference between mostly dead and all dead.” Enter Pennsylvania, where an electoral vote bill is more than slightly alive.
State Senate president Dominic Pileggi (R) is pushing a plan to change Pennsylvania’s current winner-take-all allocation of the state’s 20 electoral votes to a system that apportions them proportionally by each candidate’s share of the statewide vote. Had it gone into effect before the 2012 election, President Obama would have won just 12 electoral votes to Mitt Romney’s 8. According to a memo Pileggi wrote debuting his new plan, the revamped system “much more accurately reflects the will of the voters in our state.”
Pileggi’s bill is slightly less far-reaching a power grab than bills proposed in states like Virginia that would have divided electoral votes by congressional district, which thanks to gerrymandering strongly favors Republicans even in states Obama won handily. This is by design: Pileggi pitched a congressional district split along the lines of the Virginia bill in 2011 only to face near-unanimous opposition from Republican members of Congress in the state who feared Democrats would pour millions of dollars into winning their individual districts. While Pileggi has shifted on the issue, a pair of Pennsylvania state representatives have also introduced a congressional split version since the election.
But the biggest difference that sets Pennsylvania apart is that each of the top leaders necessary to pass a bill have expressed support for changing its electoral college. Gov. Tom Corbett (R) strongly backed Pileggi’s 2011 effort. So did state House Majority Leader Mike Turzai (R), best known nationally for bragging that voter ID restrictions opposed by civil rights groups would help Romney win the state in 2012. By contrast, similar electoral vote bills elsewhere hit a wall this year after Republican statehouse leaders, key lawmakers, or governors indicated their opposition. Source I'm actually ok with this. I don't like the people that support it and why they support it, but splitting the EC based on popular vote makes sense these days. It does make sense, but only if all the states do it imo, and if they aren't under the effect of partisan gerrymandering. Imagine if only blue states did this or vice-versa, the other party would never win an election. I only agree with it if it is done by a statewide or nationwide popular vote, and not by districts. States can't really be gerrymandered. Also, it's one of those things that both sides would be pressured to do once a solid number went that way. Finally, we don't even know how it would affect an election since spending in states would be drastically altered as some ceased to be "swing states." I'm just saying, if we're really considering this, it would be nice if a lot of states or the entire country adopted it at once, as not to skew results one way or the other. If California decided to do it and only California, suddenly Democrats would no longer be guaranteed all of those 53 electoral college votes, so it gives them a bit of a disadvantage. Likewise, if Texas were to implement it, then Republicans would lose a lot of their guaranteed electoral college votes, so I think if we were to implement this, it should be widely encompassing. That would be ideal, but nearly impossible from a practical point of view. The best I would hope for is that 3 or 4 smaller states would do it every 1-2 election cycles, until the bigger states (like Texas and California) wouldn't feel they would cost somebody an election. Even in the case of those big states, however, they are all strongly leaning one way or another. At most, we're talking about a loss of 30%, so about 16 votes from California and 12 from Texas.
Are you sure at most were talking a loss of 30%, I haven't looked at polling results etc, but I would imagine Democrats would be more likely to vote under a different system in Texas, same for Republicans in California. Again, I haven't seen the numbers from recent elections, but I don't think Texas is as Red as you think, they actually had a lot of Blue Districts not that long ago, and I would imagine California is not as liberal as you think, they didn't even get the popular vote on gay marriage not so long ago.
I'm kind of leaning towards if one of those states changed and the other didn't you would see a pretty big problem for one of the political parties.
|
Well, there is the http://en.wikipedia.org/wiki/National_Popular_Vote_Interstate_Compact Basically, a law that only comes into effect once at least half of the electoral votes worth of states have enacted it.
If that happens, all of their electoral votes from each state that enacted it, then go to the winner of the national popular vote. And since it only goes into effect once the majority of electoral votes agrees to it, it would effectively mean the president would then be elected by direct national vote, and effectively replacing the whole electoral college system.
|
On February 19 2013 10:41 sam!zdat wrote:
No, go read Taleb's book.
Not sure wich book of him you mean.
Antifragile seems to be his most recent book but it does not seem to deal with probability of the stockmarket at all.
Thx for mentioning him, i didnt knew him and he has lots of interesting publications.
O ty,hmm that name is kinda famour i didnt knew the term black sawn originated from him.
Am reading about fragile and antifragile now wich is quiet interesting.
The black swan,and another work of him (no, small probabilitys are not overpriced) basicly say that the effects of these small probabilitys can be huge, and also that the exact probability itself can not be determined due to a lack of data (in the case of the stockmarket options)
It is indeed a bit dangerous to take the bellcurve and simply extrapolate, specially for the tail ends as he mentions.
The exact change can not be known,(and is underestimated according to him) so its impossible to make calculations with it, but what is known is that the change on a huge fluctuation is verry verry small, and i still think that the fluctuations of the stockmarket do more or less follow a bell curve ,
Annyway thx for mentioning him, i learned a few new things wich i underestimated slightly
|
The Black Swan
He will explain why you should Beware the Gaussian
|
|
|
So women should get the vote in 2080ish...
|
Interesting...
The output effect of fiscal consolidations
August 2012
Abstract
This paper studies whether fiscal corrections cause large output losses. We find that it matters crucially how the fiscal correction occurs. Adjustments based upon spending cuts are much less costly in terms of output losses than tax-based ones. Spending-based adjustments have been associated with mild and short-lived recessions, in many cases with no recession at all. Tax-based adjustments have been associated with prolonged and deep recessions. The difference cannot be explained by different monetary policies during the two types of adjustments. Studying the effects of multi-year fiscal plans rather than individual shifts in fiscal variables we make progress on question of anticipated versus unanticipated policy shifts: we find that the correlation between unanticipated and anticipated shifts in taxes and spending is heterogenous across countries, suggesting that the degree of persistence of fiscal corrections varies..Estimating the effects of fiscal plans, rather than individual fiscal shocks, we obtain much more precise estimates of tax and spending multipliers.
Link
|
On February 18 2013 04:21 sam!zdat wrote:Show nested quote +On February 17 2013 19:01 Danglars wrote:On February 17 2013 16:17 sam!zdat wrote:
edit: sorry, I'm really confused. you're going to have to explain how you interpret this exhibit. of course it wasn't good DURING the war The Keynesian theory that an increase government spending is the way out of a depression or lessen the effects of a depression. Wars incur massive government spending. So many economists (google would turn up most economists) argue that the spending incurred from WW2 was huge in getting us out of the great depression. If only we could do this kind of spending in every depression! What JonnyBNoHo has there is the disgusting truth about the whole story of what happens. We ration common goods like sugar, and you can't even go where you want in your car. Government sells its bonds on patriotism, and gives kickbacks, (but they aren't selling on their own merits!). If you would like the summary in rap form, hear this one out in 30 seconds. Some call WWII spending a case in point, but it comes with nasty side effects, perhaps significant to outweigh the benefits. Yeah, I get it... You realize Keynesianism is just bandaid Marxism, right? Keynes just rediscovered what Marx already knew and pretended he didn't know about Marx so the west would accept it (Keynes did not discover the "crisis of effective demand"). Only Keynes decided the best thing to do would be to shovel more babies into Moloch. good going, Keynes. So yeah, the people suffer so that we can take away their livelihood to restart the military-industrial machine. Is any of this supposed to be a problem for me? Obviously it wasn't good for SOCIETY, but it was good for CAPITALISM.
I take it you understood the exhibit, but are moving on logically. I don't think you realize I only wanted to put in context the relevance of JonnyBNoHo's picture in a discussion that brought up the WWII and depression lessons topic. I made no societal comparisons, nor intended to. I made no broad reflections on capitalism as a whole, nor intended to. I say only that WWII might be believed to have done one thing but at a cost.
|
On February 18 2013 12:08 sam!zdat wrote:
Yeah, I get it Jonny. the point is that you get these more complicated and more complicated models to try to express risk quantitatively using seriously flawed mathematical tools (the Gaussian) and the more complicated your air castles get, the more you are deluding yourself into thinking you understand what is going on.
You should make a difference between the tool, and its (sometimes misapplied) usage. The normal distribution in itself is not flawed, and in fact incredibly beautiful.
To me the worrying part is the design of the whole economic framework rather than the interpretations thereof. You could use any non-parametric techniques to gain understanding of the current model, but if the model itself is not only incomplete (that is a given and not a serious objection) but actually a representation of a system that is toxic, the results will also be thus.
|
Cayman Islands24199 Posts
On February 19 2013 15:22 sam!zdat wrote:The Black SwanHe will explain why you should Beware the Gaussian
i don't think you should treat taleb like a god. the gaussian-mandelbrot thing is just a choice way of representation, more generalized ways of representing the same basic problem have been made elsewhere.
at the end of the day, it's simply this. big chunk of stuff controlled by a group of people whose behavior are interdependent.
|
Cayman Islands24199 Posts
On February 19 2013 17:03 JonnyBNoHo wrote:Interesting... Show nested quote +The output effect of fiscal consolidations
August 2012
Abstract
This paper studies whether fiscal corrections cause large output losses. We find that it matters crucially how the fiscal correction occurs. Adjustments based upon spending cuts are much less costly in terms of output losses than tax-based ones. Spending-based adjustments have been associated with mild and short-lived recessions, in many cases with no recession at all. Tax-based adjustments have been associated with prolonged and deep recessions. The difference cannot be explained by different monetary policies during the two types of adjustments. Studying the effects of multi-year fiscal plans rather than individual shifts in fiscal variables we make progress on question of anticipated versus unanticipated policy shifts: we find that the correlation between unanticipated and anticipated shifts in taxes and spending is heterogenous across countries, suggesting that the degree of persistence of fiscal corrections varies..Estimating the effects of fiscal plans, rather than individual fiscal shocks, we obtain much more precise estimates of tax and spending multipliers. Link
didn't read this in detail but i do buy the psychological impact theory of taxation given how businsss people talk everyday. though, this does not mean that their reaction is at all rational.
the empirical study part is the key to the paper as always
|
On February 19 2013 17:03 JonnyBNoHo wrote:Interesting... Show nested quote +The output effect of fiscal consolidations
August 2012
Abstract
This paper studies whether fiscal corrections cause large output losses. We find that it matters crucially how the fiscal correction occurs. Adjustments based upon spending cuts are much less costly in terms of output losses than tax-based ones. Spending-based adjustments have been associated with mild and short-lived recessions, in many cases with no recession at all. Tax-based adjustments have been associated with prolonged and deep recessions. The difference cannot be explained by different monetary policies during the two types of adjustments. Studying the effects of multi-year fiscal plans rather than individual shifts in fiscal variables we make progress on question of anticipated versus unanticipated policy shifts: we find that the correlation between unanticipated and anticipated shifts in taxes and spending is heterogenous across countries, suggesting that the degree of persistence of fiscal corrections varies..Estimating the effects of fiscal plans, rather than individual fiscal shocks, we obtain much more precise estimates of tax and spending multipliers. Link
I've seen that paper before, it was mentioned on Mankiw's blog, although I haven't gone through it's details.
Alberto Alesina strikes again.
That name might sound slightly familiar. The reason is because Alesina is the father of "Expansionary Austerity", the man who convinced Europe to go all in for austerity.
This latest paper says that spending cuts reduces output more than tax increases. But it's worth noting that it's contradicted by the work of... Alesina, who shows that spending cuts doesn't reduce output, but rather increases it. Here's Alesina's paper, this is the intellectual foundation for expansionary austerity.
It's possible that Alesina is correct, that tax increases are less contractionary than spending cuts. But one reason to show some doubt is that if this were true, we should expect a good economic recovery in the UK. After all, their austerity program has focused on spending cuts, while cutting taxes. They've increased the tax free threshold for income tax and reduced the corporate tax rate to amongst the lowest in the world.
|
This whole debate about using normal distributions in finance seems completely misplaced.
Firstly, the main model, Black-Scholes, doesn't assume normality. It assumes that changes in stock prices are lognormally distributed, i.e. stock prices follow a geometric Brownian motion (not a Brownian motion).
But the point is taken that these models do not allow for tail events (unpredictably huge and rare events). One above poster suggests using nonparametric methods. But using a nonparametric method means not assuming a distribution by having the distribution being dictated by the data. It doesn't mean the model is robust under any distribution. Because nonparametric methods rely on data, and there is by definition, virtually no data on tail events, nonparametric methods won't solve your problem. You can't create information ex nihilo.
So what should you do about tail events? Hedge. Buy and sell derivatives to cover your ass. Do scenario analysis, i.e. what happens to your portfolio when some big bad and unpredictable event happens?
But weren't derivatives and financial instruments that no one really understood one of the main causes of the financial crisis? Yes, but the problem wasn't that people didn't price risk properly because they used the normal distribution, it was that people didn't see the correlation and interconnectedness of the whole system. You hedge to reduce risk, but when everything goes wrong at once and the assets you have become worthless and those who owe you because of your hedge also have worthless assets, then they can't pay you and you go broke (or get a bail out).
For example, the rationale behind subprime loans (which were individually worthless because they're likely to default) was that they can be all packaged up into one entity called a MBS. The MBS was highly rated by rating agencies, because what's the chance that they'll all default? This is basically risk pooling, the idea behind insurance. But it doesn't work when things are correlated.
The problem wasn't the normal distribution. It was correlation.
To be fair, it clearly wasn't easy to model correlation or to understand that it could be a big deal. They also had intellectual cover as finance was enthralled to free market fundamentalism, the centerpiece being the Efficient Market Hypothesis (EMH). And it was argued by many people of influence, like Alan Greenspan and Larry Summers that financial innovation had dispersed risk and that because of the EMH and Chicago School of Economics that nothing could go wrong. Clearly, that didn't work out.
|
On February 19 2013 19:09 Danglars wrote:
I don't think you realize I only wanted to put in context the relevance of JonnyBNoHo's picture in a discussion that brought up the WWII and depression lessons topic. I made no societal comparisons, nor intended to. I made no broad reflections on capitalism as a whole, nor intended to. I say only that WWII might be believed to have done one thing but at a cost.
[/url]
Cool, then we're on the same page.
On February 19 2013 20:33 Fwmeh wrote:Show nested quote +On February 18 2013 12:08 sam!zdat wrote:
Yeah, I get it Jonny. the point is that you get these more complicated and more complicated models to try to express risk quantitatively using seriously flawed mathematical tools (the Gaussian) and the more complicated your air castles get, the more you are deluding yourself into thinking you understand what is going on. You should make a difference between the tool, and its (sometimes misapplied) usage. The normal distribution in itself is not flawed, and in fact incredibly beautiful.
Er, ok, it's flawed when you want to use it for studying things which are in fact mandelbrotian and not gaussian. I'll leave any appreciation of the platonic form of the bell curve to others, it's fine for what it is. That's not really my point.
To me the worrying part is the design of the whole economic framework rather than the interpretations thereof. You could use any non-parametric techniques to gain understanding of the current model, but if the model itself is not only incomplete (that is a given and not a serious objection) but actually a representation of a system that is toxic, the results will also be thus.
I'm worried about both, so yes I agree that we shouldn't just get preoccupied with the fact that we're modeling it badly, because yes, the thing we're trying to model is already a bad idea.
On February 19 2013 21:43 oneofthem wrote:Show nested quote +On February 19 2013 15:22 sam!zdat wrote:The Black SwanHe will explain why you should Beware the Gaussian i don't think you should treat taleb like a god. the gaussian-mandelbrot thing is just a choice way of representation, more generalized ways of representing the same basic problem have been made elsewhere.
How could I treat like a God anybody who says nice things about Hayek? I just want people to read Taleb, because he's somebody who's going to convince people of this more than I, because he has background in finance and he's not also trying to convince them to become communists at the same time. If I can teach some suits some epistemic humility, I'll count that as a small blow against Empire. And Taleb seems like the man for the job. What else should I recommend to people?
I might be guilty of deifying mandelbrot, but that is an entirely different topic 
at the end of the day, it's simply this. big chunk of stuff controlled by a group of people whose behavior are interdependent.
hmm. I feel like that's what it is at the beginning of the day. Then you have to go theorize it.
On February 20 2013 00:37 paralleluniverse wrote:
But the point is taken that these models do not allow for tail events (unpredictably huge and rare events).
The point is that this is what we call History.
The problem wasn't the normal distribution. It was correlation.
These two are related. When you are using something like a normal distribution to measure something, you are assuming that the things you are measuring are uncorrelated. "the bell curve" is just short of a short hand for "models which assume that there will not be any tail events"
To be fair, it clearly wasn't easy to model correlation or to understand that it could be a big deal. They also had intellectual cover as finance was enthralled to free market fundamentalism, the centerpiece being the Efficient Market Hypothesis (EMH). And it was argued by many people of influence, like Alan Greenspan and Larry Summers that financial innovation had dispersed risk and that because of the EMH and Chicago School of Economics that nothing could go wrong. Clearly, that didn't work out.
I don't think free market fundamentalism is an excuse for anyone to do anything at all. Nothing "intellectual" about free market fundamentalism. I think a five year old could have told them that correlation would be a big deal. Anybody tells you "oh, yeah, we solved History, don't worry, nothing can go wrong now because we have math that tells us everything" you punch them in the face. The fact that anybody ever took these people seriously does not exactly help my opinion of financiers and economists.
On February 20 2013 00:37 paralleluniverse wrote:
So what should you do about tail events? Hedge. Buy and sell derivatives to cover your ass. Do scenario analysis, i.e. what happens to your portfolio when some big bad and unpredictable event happens?
No, I think this is missing the point. You should create a robust system which is multiply redundant and de-globalized, and use fewer financial tools in general. The derivates can't help you, because the derivates also don't know anything about what they don't know.
YOU CANNOT DO ANALYSIS FOR SOMETHING BIG BAD AND UNPREDICTABLE. If you do analysis for it, it is just big bad and predictable. Then it becomes a "gray swan." You can do this, sure, that's part of the point. Try to turn black swans into gray swans. But never think that there aren't more black swans out there.
|
On February 20 2013 00:37 paralleluniverse wrote:
This whole debate about using normal distributions in finance seems completely misplaced.
Firstly, the main model, Black-Scholes, doesn't assume normality. It assumes that changes in stock prices are lognormally distributed, i.e. stock prices follow a geometric Brownian motion (not a Brownian motion).
But the point is taken that these models do not allow for tail events (unpredictably huge and rare events). One above poster suggests using nonparametric methods. But using a nonparametric method means not assuming a distribution by having the distribution being dictated by the data. It doesn't mean the model is robust under any distribution. Because nonparametric methods rely on data, and there is by definition, virtually no data on tail events, nonparametric methods won't solve your problem. You can't create information ex nihilo.
So what should you do about tail events? Hedge. Buy and sell derivatives to cover your ass. Do scenario analysis, i.e. what happens to your portfolio when some big bad and unpredictable event happens?
But weren't derivatives and financial instruments that no one really understood one of the main causes of the financial crisis? Yes, but the problem wasn't that people didn't price risk properly because they used the normal distribution, it was that people didn't see the correlation and interconnectedness of the whole system. You hedge to reduce risk, but when everything goes wrong at once and the assets you have become worthless and those who owe you because of your hedge also have worthless assets, then they can't pay you and you go broke (or get a bail out).
For example, the rationale behind subprime loans (which were individually worthless because they're likely to default) was that they can be all packaged up into one entity called a MBS. The MBS was highly rated by rating agencies, because what's the chance that they'll all default? This is basically risk pooling, the idea behind insurance. But it doesn't work when things are correlated.
The problem wasn't the normal distribution. It was correlation.
To be fair, it clearly wasn't easy to model correlation or to understand that it could be a big deal. They also had intellectual cover as finance was enthralled to free market fundamentalism, the centerpiece being the Efficient Market Hypothesis (EMH). And it was argued by many people of influence, like Alan Greenspan and Larry Summers that financial innovation had dispersed risk and that because of the EMH and Chicago School of Economics that nothing could go wrong. Clearly, that didn't work out.
Well normal distributions do have tails, and thus tail risk, so there's an argument to be made that the tails just weren't thick enough.
There's also evidence that the tail risk show was simply ignored. AAA tranches seem to have been fairly priced (extremely low default rates) + Show Spoiler + but many of the lower grade tranches were clearly overpriced. If people ignore the risk then it really doesn't matter what distribution you use.
As far as hedging goes it was a mixed bag. Generic portfolio diversification didn't work (correlation problem) but other more direct hedges, like derivatives, worked as intended and only failed if the counter party was unable to meet his obligations (counter party risk).
|
Cayman Islands24199 Posts
of course, the large US Govt counterparty was there...
|
On February 20 2013 03:29 JonnyBNoHo wrote:
Well normal distributions do have tails, and thus tail risk, so there's an argument to be made that the tails just weren't thick enough.
No, the tails are of entirely the wrong structure.
There's also evidence that the tail risk show was simply ignored. AAA tranches seem to have been fairly priced (extremely low default rates) + Show Spoiler + but many of the lower grade tranches were clearly overpriced. If people ignore the risk then it really doesn't matter what distribution you use.
Sure, we can't blame bad models for everything. The proximal causes of the crash were greed, lies, and corruption. The bad models just provided some accoutrements for the priesthood to wave in our faces.
As far as hedging goes it was a mixed bag. Generic portfolio diversification didn't work (correlation problem) but other more direct hedges, like derivatives, worked as intended and only failed if the counter party was unable to meet his obligations (counter party risk).
Someday the financiers will discover poststructuralism, and this will be the only useful thing poststructuralism ever accomplished. Where is the structuring center? where is the counter party which is counter to all other parties? is it God? Freedom? Adam Smith's rotting corpse?
|
|
|
|
|
|
EPT![[image loading]](http://www.standardandpoors.com/spf/fgr_articles/950068/6750010.gif)
