DHT,
I have always been a proponent of that the bulk of your assets should be denominated in the currency of the country where you plan to depend on your investment assets. Have had many discussions and debates on this. The only reasonable counter (in my opinion) that I have heard to this is runaway inflation (hyper inflation) in the country of investment.
I have heard arguments using the MPT and diversification by asset classes and different currency being different classes - though I do not buy that as far as currency is concerned. The principle of interest rate parity says that exchange rate gets compensated by interest rate differential but in reality a lot more including controls by central banks go into it. While investing in different asset classes lowers risk, I do not buy that being invested in different currencies lowers risk - I think it increases it. Hyperinflation is the only risk that I would agree on there are counters to that.
Correlation of Assets and its effect in Asset Allocation
Correlation of Assets and its effect in Asset Allocation
Desi;51528
I have always been a proponent of that the bulk of your assets should be denominated in the currency of the country where you plan to depend on your investment assets. Have had many discussions and debates on this. The only reasonable counter (in my opinion) that I have heard to this is runaway inflation (hyper inflation) in the country of investment.
[/quote]
For fixed income portion, I would suggest 100% in local debt market and deviation only in special tax circumstances.
However, for equity portion, I would suggest diversification across the world, with some slight home country bias - with the proportion to be invested in home country depending on relative size of home country equity market and stability to that of world equity markets.
For example, consider Ram Bahadur Thapa who wants to retire in Kathmandu. Of the equity portion of his AAP - perhaps at most 10% is the appropriate investment in Nepalese equity market coz Nepal is small. India equity market is bigger, yet small relative to the size of the world market.
Correlation of Assets and its effect in Asset Allocation
Yes, when we consider countries such as Nepal or Mongolia, we are hard pressed to find public equity markets. To participate in business growth in such countries becomes impossible till one establishes one's own business or partners in a business with a few others. Both of those approaches are non diversified and carry a risk. In such a case one can participate in businesses outside the country to seek equity growth. It comes with a foreign exchange risk. I know many on this forum whose knowledge I respect disagee with me and many economists of repute will disagree with me, but I have not been able to convince myself that diversifying outside the country of denomination (in established markets) lowers risk overall.
Those in markets such as USA or other reasonably developed and diversified equity markets, should they invest outside their country of denomination?
Does this investment outside the country of denomination lower the overall risk by providing diversification without sacrificing growth? Of course diversification lowers risk, but is this lowering of risk cancelled out by currency risk assumption. In my opinion the additional diversification may reduce risk much less than the risk assumed by currency translation.
Should an LIA for example invest in Europe, Asia and Latin America? It gets him investment (businesses invested in) diversification which does lower risk. But does the eventual conversion into currency of consumption give a net higher risk or a lower risk?
If the local markets are well diversified, does the additional diversification buy anything? Does additional diversification here lower risk?
I personally invest in markets outside US, but I do that as a matter of speculation that foreign markets are offering greater growth, realizing that I eventually am going to consume dollars and not Euros so have to accept a currency translation risk.
Does this currency diversification lower my risk, I think not. Is being invested in a single currency (e.g. dollar) increasing my risk, I think not. Main reason is my consumption is in dollars and by keeping money in any other currency, I have added a risk component. Personally I have even gone for emerging market bonds, but quire clearly to chase growth I take higher risks.
My speculation of investing in foreign markets paid off (so far) due to growth and dollar decline, but what about the Indian, Chinese or European investor who invested in US markets to diversify. He got the comparitively anemic US growth but suffered the dollar decline as well - the currency risk came into play.
As a matter of advice, I do advice R2Is, to have fixed income investments in India. One reason for that is that in India, even today the participation in equity markets by the average Joe is less which increases his participation in debt markets thus the returns in debt markets have a less of a discount from equity markets and provide a comparitively higher real rate of return that debt markets say in USA.
For equity markets, India to some extent poses a problem like investments in Nepal, while not exactly the same problem but somewhat similar and that has to do with transparency and regulation. This creates a dilemma as to whether the equity investment should go into less transparent markets with less regulatory oversight and greater chances of being swindled - the choice to overcome this risk is to diversify equity outside the country.
What should investors in countries such as in Europe and in USA do? Should they invest in markets outside their country? Are they assuming a greater risk? Are they chasing growth in other words, speculating growth elsewhere?
Those in markets such as USA or other reasonably developed and diversified equity markets, should they invest outside their country of denomination?
Does this investment outside the country of denomination lower the overall risk by providing diversification without sacrificing growth? Of course diversification lowers risk, but is this lowering of risk cancelled out by currency risk assumption. In my opinion the additional diversification may reduce risk much less than the risk assumed by currency translation.
Should an LIA for example invest in Europe, Asia and Latin America? It gets him investment (businesses invested in) diversification which does lower risk. But does the eventual conversion into currency of consumption give a net higher risk or a lower risk?
If the local markets are well diversified, does the additional diversification buy anything? Does additional diversification here lower risk?
I personally invest in markets outside US, but I do that as a matter of speculation that foreign markets are offering greater growth, realizing that I eventually am going to consume dollars and not Euros so have to accept a currency translation risk.
Does this currency diversification lower my risk, I think not. Is being invested in a single currency (e.g. dollar) increasing my risk, I think not. Main reason is my consumption is in dollars and by keeping money in any other currency, I have added a risk component. Personally I have even gone for emerging market bonds, but quire clearly to chase growth I take higher risks.
My speculation of investing in foreign markets paid off (so far) due to growth and dollar decline, but what about the Indian, Chinese or European investor who invested in US markets to diversify. He got the comparitively anemic US growth but suffered the dollar decline as well - the currency risk came into play.
As a matter of advice, I do advice R2Is, to have fixed income investments in India. One reason for that is that in India, even today the participation in equity markets by the average Joe is less which increases his participation in debt markets thus the returns in debt markets have a less of a discount from equity markets and provide a comparitively higher real rate of return that debt markets say in USA.
For equity markets, India to some extent poses a problem like investments in Nepal, while not exactly the same problem but somewhat similar and that has to do with transparency and regulation. This creates a dilemma as to whether the equity investment should go into less transparent markets with less regulatory oversight and greater chances of being swindled - the choice to overcome this risk is to diversify equity outside the country.
What should investors in countries such as in Europe and in USA do? Should they invest in markets outside their country? Are they assuming a greater risk? Are they chasing growth in other words, speculating growth elsewhere?
Correlation of Assets and its effect in Asset Allocation
Desi;51541
Those in markets such as USA or other reasonably developed and diversified equity markets, should they invest outside their country of denomination?
Does this investment outside the country of denomination lower the overall risk by providing diversification without sacrificing growth? Of course diversification lowers risk, but is this lowering of risk cancelled out by currency risk assumption. In my opinion the additional diversification may reduce risk much less than the risk assumed by currency translation.
Should an LIA for example invest in Europe, Asia and Latin America? It gets him investment (businesses invested in) diversification which does lower risk. But does the eventual conversion into currency of consumption give a net higher risk or a lower risk?
If the local markets are well diversified, does the additional diversification buy anything? Does additional diversification here lower risk?
[/quote]
As long as returns from non US markets (in USD terms i.e. considering exchange rate risk), even if more risky than US market, are not perfectly correlated with that of US market, there is benefit to diversification.
To get the diversification free lunch by adding a new asset to a portfolio, it is not necessary that new asset that is added have a lower risk or a higher expected return than the existing portfolio.
The magnitude of the diversification benefit is higher when the return from the new asset has lower (or better still more negative) correlation with the return from existing portfolio. With globalization these correlations are increasing and thus lowering the diversification benefit, but since the correlations are not perfect the diversification benefit is not eliminated.
The diversification benefit is also larger when the new asset class has higher expected return and/or lower risk and when the proportion of the investor's consumption that is met via imports is higher.
Correlation of Assets and its effect in Asset Allocation
When we talk of corelations, we focus on the short term corelations, in that two assets A and B negatively corelated would react oppositely to market events or positive corelation and the number shows the extent. On a stock such as VMW with high beta shoots up 5% for 1% rise in sp500.
Long term, all assets should exhibit a growth whcih is a long term positive corelation.
In fact to lower risk, the AAP should as much of uncorelated assets rather than negatively corelated assets and when that is not possible, then the negatively corelated assets. This lowers the volatility and hence the risk.
Now here is where the this corelation does not sit well with me. One - the corelation is empirical based on historical data and two - a bunch of short term negative corelatiosn should add up to one large long term negative corelation and that would mean lower all growth. That does not happen and not what we want. We want long term overall growth and all assets to show growth long term - a positive corelation.
So overall, I take this business of corelation with a grain of salt.
Long term, all assets should exhibit a growth whcih is a long term positive corelation.
In fact to lower risk, the AAP should as much of uncorelated assets rather than negatively corelated assets and when that is not possible, then the negatively corelated assets. This lowers the volatility and hence the risk.
Now here is where the this corelation does not sit well with me. One - the corelation is empirical based on historical data and two - a bunch of short term negative corelatiosn should add up to one large long term negative corelation and that would mean lower all growth. That does not happen and not what we want. We want long term overall growth and all assets to show growth long term - a positive corelation.
So overall, I take this business of corelation with a grain of salt.
Correlation of Assets and its effect in Asset Allocation
Desi;51558
Long term, all assets should exhibit a growth whcih is a long term positive corelation.
In fact to lower risk, the AAP should as much of uncorelated assets rather than negatively corelated assets and when that is not possible, then the negatively corelated assets. This lowers the volatility and hence the risk.
[/quote]
Looks like you are making the following statements:
(i) Given two assets A and B, if both assets have positive expected return, they cannot possibly have a negative correlation.
(ii) Given two assets C and D, if the two assets are negatively correlated and asset C has positive expected return, then asset D must necessarily have negative expected return.
If so, I submit that both statements (i) and (ii) above are incorrect.
It is not uncommon to see people who believe the two statements are correct - usually their understanding stems from viewing negative correlation between two assets solely in prosaic terms as "If one asset's price goes up, then the price of the other goes down" and not going any further. If that is the entire view one has of negative correlation, then it is a misleading view of what constitutes negative correlation. It is possible for two assets E and F to exhibit positive returns in every year for 30 years and still have their returns be negatively correlated over the same 30 year period.
Now negatively correlated assets are difficult to find - however that is different from saying that they are not desirable.
As far as known correlations being dependent on historical info is concerned - well, historical info is all we have to predict these parameters, even as we know that these parameters can change.
Correlation of Assets and its effect in Asset Allocation
Bobus, #55 is great explanation on correlation.
Desi,
While I agree in principle, that one should have assets in currency of domicile, other factors need consideration too.
You pointed out that in #51 as below:
[QUOTE]For equity markets, India to some extent poses a problem like investments in Nepal, while not exactly the same problem but somewhat similar and that has to do with transparency and regulation. This creates a dilemma as to whether the equity investment should go into less transparent markets with less regulatory oversight and greater chances of being swindled - the choice to overcome this risk is to diversify equity outside the country.
The answer lies in the above para.
Emerging markets pose some serious problems related to regulations, fraud, accounting manipulations etc., An investor need to tackle all these problems in addition to volatality. Volatality is a risk in short term and the former is an issue for long term investing. This is why I would say an investor need more equity exposure in developed markets than emerging. For example, does India have mechanism to prevent an enron or worldcom ? For that matter, do we have accounting regulations to identify such issues ?
Currency risk is minimal compared to these risks, in a global economy where all countries wish to export more of their products out.
Desi,
While I agree in principle, that one should have assets in currency of domicile, other factors need consideration too.
You pointed out that in #51 as below:
[QUOTE]For equity markets, India to some extent poses a problem like investments in Nepal, while not exactly the same problem but somewhat similar and that has to do with transparency and regulation. This creates a dilemma as to whether the equity investment should go into less transparent markets with less regulatory oversight and greater chances of being swindled - the choice to overcome this risk is to diversify equity outside the country.
The answer lies in the above para.
Emerging markets pose some serious problems related to regulations, fraud, accounting manipulations etc., An investor need to tackle all these problems in addition to volatality. Volatality is a risk in short term and the former is an issue for long term investing. This is why I would say an investor need more equity exposure in developed markets than emerging. For example, does India have mechanism to prevent an enron or worldcom ? For that matter, do we have accounting regulations to identify such issues ?
Currency risk is minimal compared to these risks, in a global economy where all countries wish to export more of their products out.
Correlation of Assets and its effect in Asset Allocation
Bobus;51569Looks like you are making the following statements:
(i) Given two assets A and B, if both assets have positive expected return, they cannot possibly have a negative correlation.
(ii) Given two assets C and D, if the two assets are negatively correlated and asset C has positive expected return, then asset D must necessarily have negative expected return.
If so, I submit that both statements (i) and (ii) above are incorrect. [/quote]
Bobus,
here is what I said "Now here is where the this corelation does not sit well with me. One - the corelation is empirical based on historical data and two - a bunch of short term negative corelatiosn should add up to one large long term negative corelation and that would mean lower all growth. "
I do realize for example that bonds and stocks both are positively corelated over the long term and over short term negatively corelated. I understand the reasons for it also and principally the bond growth comes from interest.
But help me understand besides bonds, if an asset has a negative corelation in short term, why would it have positive corelation in long term. I have not studied the methods used to compute the corelation, if you have a link or a book reference that would be appreciated.
[QUOTE]It is not uncommon to see people who believe the two statements are correct - usually their understanding stems from viewing negative correlation between two assets solely in prosaic terms as "If one asset's price goes up, then the price of the other goes down" and not going any further. If that is the entire view one has of negative correlation, then it is a misleading view of what constitutes negative correlation. It is possible for two assets E and F to exhibit positive returns in every year for 30 years and still have their returns be negatively correlated over the same 30 year period.
If you could expound further on that or provide some reading material references, that would be greatly appreciated.
[QUOTE]Now negatively correlated assets are difficult to find - however that is different from saying that they are not desirable.
Actually, if one can call shorting an asset then there are many ETFs and open ended short funds, but these will exhibit true short term negative corelation as well as long term negative corelation.
Comparison of SPY to SH
Obviously, we do not want such negatively corelated assets (SPY and SH) in a portfolio as they will provide a zero or a net negative return long term.
So a negative corelation short term but a positive corelation long term is what is needed in a portfolio. Negative corelation to lower net volatility and positive long term corelation for net long term growth.
[QUOTE]
As far as known correlations being dependent on historical info is concerned - well, historical info is all we have to predict these parameters, even as we know that these parameters can change.
Yes. I agree. I am not contesting that there needs to be a better method or a different method - this is all we got. My heartburn stems from reliability of such data or more correctly the weight such empirical data should be given during portfolio construction. For me, if I see a negative corelation or no corelation, then I would like to peel the layers to see inside the cause.
Correlation of Assets and its effect in Asset Allocation
Desi:
The statements
So a negative corelation short term but a positive corelation long term is what is needed in a portfolio. Negative corelation to lower net volatility and positive long term corelation for net long term growth.
and
Long term, all assets should exhibit a growth whcih is a long term positive corelation.
are erroneous and suggest that you are confusing positive expected return and positive correlation. What is needed is positive expected return, not positive correlation.
I believe one step at a time will likely be more fruitful. For starters:
I suggest taking a look at the section on negative correlation and Table 2 on Page 2 of the link below. Though the data is hypothetical (the two assets are perfectly negatively correlated), it can clear some misconceptions (though possibly not all**) about what constitutes negative correlation. In Table 2 of the link below, both assets have positive return of 10% (over the 6 year window) and still are perfectly negatively correlated - so it can help to dispel the notion that assets that have positive return cannot be negatively correlated.
http://www.tiaa-cref.org/administrators/pdf/admin_library/C36375.pdf
** The misconception that the example in Table 2 of the link above may not clear is that when two assets are negatively correlated, if one asset has positive return the other asset shd have a negative return. The example in Table 2 has the feature that when one has positive return, the other has negative return - yet even this feature is not necessary for two assets to exhibit negative correlation.
If two assets A and B are negative correlated, it just means that when A's return is above A's mean return (expected return), then B's return tends (not necessary for this to occur every period) to be below B's mean return, and vice-versa. The mean return of both A and B can be positive. Also, return in every period can be positive for both A and B. When A's return goes up (above its mean), then B's return tends to go down (below B's mean) - however B's return need not necessarily go negative (below zero) for the return from the two assets to be negatively correlated.
Please feel free to let me know if you have questions / disagreements about the above or seek elaboration. After we on the same page about the above, we can move to the other issues, which are related.
The statements
So a negative corelation short term but a positive corelation long term is what is needed in a portfolio. Negative corelation to lower net volatility and positive long term corelation for net long term growth.
and
Long term, all assets should exhibit a growth whcih is a long term positive corelation.
are erroneous and suggest that you are confusing positive expected return and positive correlation. What is needed is positive expected return, not positive correlation.
I believe one step at a time will likely be more fruitful. For starters:
I suggest taking a look at the section on negative correlation and Table 2 on Page 2 of the link below. Though the data is hypothetical (the two assets are perfectly negatively correlated), it can clear some misconceptions (though possibly not all**) about what constitutes negative correlation. In Table 2 of the link below, both assets have positive return of 10% (over the 6 year window) and still are perfectly negatively correlated - so it can help to dispel the notion that assets that have positive return cannot be negatively correlated.
http://www.tiaa-cref.org/administrators/pdf/admin_library/C36375.pdf
** The misconception that the example in Table 2 of the link above may not clear is that when two assets are negatively correlated, if one asset has positive return the other asset shd have a negative return. The example in Table 2 has the feature that when one has positive return, the other has negative return - yet even this feature is not necessary for two assets to exhibit negative correlation.
If two assets A and B are negative correlated, it just means that when A's return is above A's mean return (expected return), then B's return tends (not necessary for this to occur every period) to be below B's mean return, and vice-versa. The mean return of both A and B can be positive. Also, return in every period can be positive for both A and B. When A's return goes up (above its mean), then B's return tends to go down (below B's mean) - however B's return need not necessarily go negative (below zero) for the return from the two assets to be negatively correlated.
Please feel free to let me know if you have questions / disagreements about the above or seek elaboration. After we on the same page about the above, we can move to the other issues, which are related.
Correlation of Assets and its effect in Asset Allocation
Bobus,
Thanks, I will go thru the link and post further questions later. Agreed that positive corelation and positve expected return are two different things. Two assets A and B can both exhibit a negative return say in in year 2008 and both will have a positive corelation.
Long term, all assets should exhibit a growth whcih is a long term positive corelation.
I worded my statement poorly. What I meant by above statement that long term I want those assets that have a positive expected return. All my assets should have a long term positive expected return. Whether they give me an actual positive or negative return will depend on realities of the economy and even though the long term expected return is positive, I may get a net negative return (Japanese market).
Now if the assets A and B have long term expected positive return, should not their corelation be positive?
So how is it possible to have negative corelation short term while having positive corelation long term.
I will re read your posts a couple of more times as well as the TIAA link and post more questions later, except one for now and that is - do you have any reference regarding the detail of methods and mechanisms used to compute beta (in relationship to SP500)?
Thanks, I will go thru the link and post further questions later. Agreed that positive corelation and positve expected return are two different things. Two assets A and B can both exhibit a negative return say in in year 2008 and both will have a positive corelation.
Long term, all assets should exhibit a growth whcih is a long term positive corelation.
I worded my statement poorly. What I meant by above statement that long term I want those assets that have a positive expected return. All my assets should have a long term positive expected return. Whether they give me an actual positive or negative return will depend on realities of the economy and even though the long term expected return is positive, I may get a net negative return (Japanese market).
Now if the assets A and B have long term expected positive return, should not their corelation be positive?
So how is it possible to have negative corelation short term while having positive corelation long term.
I will re read your posts a couple of more times as well as the TIAA link and post more questions later, except one for now and that is - do you have any reference regarding the detail of methods and mechanisms used to compute beta (in relationship to SP500)?