Taleb talking about his portfolio strategy. In the 90% in 'safe' assets, he owns a diversified pool of inflation linked bonds in different currencies along with gold (I believe) and other metals. That's a very 'balanced' type approach, just designed to have less returns and less fluctuations
Global Macro Trading Journal
- Thread starter Daal
- Start date
This Taleb approach is something I had thought about in the past. I thought about it once I was thinking about the IMF's "drawing rights". These drawing rights are a unit of account of wealth that involve a basket of currencies. I don't necessarily agree with the weightings and the components of this unit of accout. For instance, over the last few years it was worth
A better approach is probably select a pool of currencies that stood the test of time (that have been around for a really long-time) and add gold and silver in it. This becomes a new way to measure "wealth". So if someone buys $100 USD worth of AAPL on day 1 'and that is worth, say, 0.56 of the new 'unit of account'). If that drops to $95, it looks like there was a 5% drop, but if measured in the new unit of account, it was worth 0.57, there was actually a gain in the trade.
Here is a mix that I think it makes sense if the goal is 100% safety:
-Swiss Franc (historically a very stable currency/country). CHF has been around since 1850
-US Dollar. 1792-Present, very stable country. Its on the Rogoff never defaulted list
-Australian Dollar. Country is on the Rogoff list of countries that never defaulted.
-Danish Krone. Been around since 1875. Its on Rogoff's list of countries that never defaulted (only one of the nordic countries)
-Gold. 5000 year history as store of value. Stood the test of time
-Silver. Same thing
There are some other possible currencies like the Canadian dollar and the Norwegian/Swedish Krone but they are somewhat correlated and close to the other ones in the unit of account (either because of commodities or because of geographical location) so they wouldn't add much in value. The EUR I kinda like (beause it if drops you can always travel to several countries and spend the money there, so it isn't a 'real' loss) but its too new and untested. The JPY was wiped out in the past, trades over 100 to the dollar for a reason, country is deep into debt. The Chinese Yuan is also new, I believe they also had a hyperinflation in the past. The GBP is a potential idea, it wouldn't be a terrible addition
In order to not try to weight things in a biased way, it would be an equal weighted measure.
So 16.66% of each.
This would be an ultimate 'store of value'. Some gold bugs put all they got into gold but that creates some significant volatility as gold can go through several years as a pretty bad store of value. A basket approach adds diversification to this and is likely to diminish fluctuations greatly (Its also close to my 15% rule of thumb of optimum allocations with a limited menu of assets)
These are the components on the last trading day of Dec 2016, measured in US dollars
AUD 1.3858
USD 1
CHF 1.0184
DKK 7.0643
Gold 1159.10
Silver 15.989
I will try to keep track of this 'store of value index' to use as a benchmark overtime. But the idea is that using a measure of wealth that is stable (that is robust) will serve the person more than using their home currency. If someone always measure their worth in their domestic currency, that person is extremely vulnerable to extreme events. The person will load up on the domestic currency to avoid FX fluctuations and this will lead to excessive concentration of risk. Most of the time it will work but if an extreme event like a lot of inflation happens (or worse, hyperinflation) the person's net worth will drop like a rock in terms of the store of value index.
A better approach is probably select a pool of currencies that stood the test of time (that have been around for a really long-time) and add gold and silver in it. This becomes a new way to measure "wealth". So if someone buys $100 USD worth of AAPL on day 1 'and that is worth, say, 0.56 of the new 'unit of account'). If that drops to $95, it looks like there was a 5% drop, but if measured in the new unit of account, it was worth 0.57, there was actually a gain in the trade.
Here is a mix that I think it makes sense if the goal is 100% safety:
-Swiss Franc (historically a very stable currency/country). CHF has been around since 1850
-US Dollar. 1792-Present, very stable country. Its on the Rogoff never defaulted list
-Australian Dollar. Country is on the Rogoff list of countries that never defaulted.
-Danish Krone. Been around since 1875. Its on Rogoff's list of countries that never defaulted (only one of the nordic countries)
-Gold. 5000 year history as store of value. Stood the test of time
-Silver. Same thing
There are some other possible currencies like the Canadian dollar and the Norwegian/Swedish Krone but they are somewhat correlated and close to the other ones in the unit of account (either because of commodities or because of geographical location) so they wouldn't add much in value. The EUR I kinda like (beause it if drops you can always travel to several countries and spend the money there, so it isn't a 'real' loss) but its too new and untested. The JPY was wiped out in the past, trades over 100 to the dollar for a reason, country is deep into debt. The Chinese Yuan is also new, I believe they also had a hyperinflation in the past. The GBP is a potential idea, it wouldn't be a terrible addition
In order to not try to weight things in a biased way, it would be an equal weighted measure.
So 16.66% of each.
This would be an ultimate 'store of value'. Some gold bugs put all they got into gold but that creates some significant volatility as gold can go through several years as a pretty bad store of value. A basket approach adds diversification to this and is likely to diminish fluctuations greatly (Its also close to my 15% rule of thumb of optimum allocations with a limited menu of assets)
These are the components on the last trading day of Dec 2016, measured in US dollars
AUD 1.3858
USD 1
CHF 1.0184
DKK 7.0643
Gold 1159.10
Silver 15.989
I will try to keep track of this 'store of value index' to use as a benchmark overtime. But the idea is that using a measure of wealth that is stable (that is robust) will serve the person more than using their home currency. If someone always measure their worth in their domestic currency, that person is extremely vulnerable to extreme events. The person will load up on the domestic currency to avoid FX fluctuations and this will lead to excessive concentration of risk. Most of the time it will work but if an extreme event like a lot of inflation happens (or worse, hyperinflation) the person's net worth will drop like a rock in terms of the store of value index.
Last edited:
Of the countries in the store of value index, only Switzerland doesn't offer inflation linked bonds but I think that's fine, its such a stable country/currency that short-term bonds/deposit rates probably will keep up with inflation over long periods of time
To measure changes in the index I will see what was the % change that year and the assume the investor earned the inflation rate (either with inflation linked bonds or by staying in cash when inflation was negative or by using short-term bonds). Except on gold and silver, where I will only take the % change into account
To measure changes in the index I will see what was the % change that year and the assume the investor earned the inflation rate (either with inflation linked bonds or by staying in cash when inflation was negative or by using short-term bonds). Except on gold and silver, where I will only take the % change into account
Last time I measured my FX exposure I was at 50% BRL 45% USD and ~5% Gold. I probably need to further diversify my USD exposure into other currencies in the index for safety and more diversification benefits. I still think it's probably optimum to have a good chunk in my domestic currency because I do have most of my costs in that currency and my country pays a lot in terms of real interest rates/real returns so it isn't hard to keep up with inflation, although I still have to give this some more thinking. But the other bucket I can probably find other places to get exposure (using derivatives)
Its tough because I dont want to take margin risk. Since my broker margins are denominated in USD, if I add a FX exposure through futures or something else that leads to some losses (measured in USD), that can lead to situation where I can run out of margins if it happens to coincide with a drawdown in my other risk assets
Its tough because I dont want to take margin risk. Since my broker margins are denominated in USD, if I add a FX exposure through futures or something else that leads to some losses (measured in USD), that can lead to situation where I can run out of margins if it happens to coincide with a drawdown in my other risk assets
One idea that might make sense is to implement a 'stop loss' approach. I got my BRL/USD/Gold exposures which I can measure against the store of value index. If I notice that I'm lagging the index and I start to get worried I'm getting devalued by those governments, I can then switch to a FX exposure that more resembles that index. I can short some USDCHF, sell some BRL futures, short USDDKK, etc. And then assume a more defensive stance of
25% BRL, 25% USD, 10% gold, 20% CHF 20% DKK. If I keep lagging the index, I can keep adjusting until I'm comfortable with the risks
25% BRL, 25% USD, 10% gold, 20% CHF 20% DKK. If I keep lagging the index, I can keep adjusting until I'm comfortable with the risks
I probably need to bring down the weighting of Silver some because its so hugely volatile. Maybe even gold a bit. Otherwise the y-y changes will be mostly dominated by them. Further improvements could be made
Inflation linked bonds are interesting investments but only if the underlying country behind them is stable. If the inflation rate is likely to fluctuatle widely, then they are not as protective as they look.
Lets say a country has a inflation linked bond promising a 2% real return with a 2% expected inflation. The tax rate is 20%. The after-tax return is 3.2% for a real return of 1.2%
But what happens if inflation soars? At 30% inflation (so the total return would be 30% plus the 2% real that was 'locked' at the start of the investment), the final return after tax will be 25.6%. Over 4% in real purchasing power loss. At 100%, the loss is over 15%. The higher the inflation rate, the higher the real loss at an accelerating rate (and that loss gets compounded over the years).
So inflation linked bonds are concave (vulnerable at an accelerating rate) to the level of inflation. So in that sense, it can be a little illusory to the buyer if they think they are truly being protected to inflation. That protection is as good as the trust you have in the currency/country.
And I'm not even factoring in that big inflation tends to lead governments to want to manipulate price indexes. So not only there will be a 'tax effect' at hurting real returns, the nominal returns will come down as inflation gets underreported (I heard that was a real problem in Argentina). That adds even more harm (concavity) to higher inflation rates. There is also the chance of outright default on the domestic debt, like Russia did in 98. Even though they could have just printed the money to pay
That's where a Brazil hedge fund manager was being contradictory when he was avoiding brazilian equities at all costs (due economic and fiscal concerns) but somehow was totally comfortable at sitting in inflation linked bonds. Equities are actually more resilient to inflation than inflation linked bonds
So emerging market and some developed market inflation linked bonds are not as safe as they look.
Out of the countries in my index, there is a lot more trust because those countries have more than 1 century of resonably stable inflation rates.
Lets say a country has a inflation linked bond promising a 2% real return with a 2% expected inflation. The tax rate is 20%. The after-tax return is 3.2% for a real return of 1.2%
But what happens if inflation soars? At 30% inflation (so the total return would be 30% plus the 2% real that was 'locked' at the start of the investment), the final return after tax will be 25.6%. Over 4% in real purchasing power loss. At 100%, the loss is over 15%. The higher the inflation rate, the higher the real loss at an accelerating rate (and that loss gets compounded over the years).
So inflation linked bonds are concave (vulnerable at an accelerating rate) to the level of inflation. So in that sense, it can be a little illusory to the buyer if they think they are truly being protected to inflation. That protection is as good as the trust you have in the currency/country.
And I'm not even factoring in that big inflation tends to lead governments to want to manipulate price indexes. So not only there will be a 'tax effect' at hurting real returns, the nominal returns will come down as inflation gets underreported (I heard that was a real problem in Argentina). That adds even more harm (concavity) to higher inflation rates. There is also the chance of outright default on the domestic debt, like Russia did in 98. Even though they could have just printed the money to pay
That's where a Brazil hedge fund manager was being contradictory when he was avoiding brazilian equities at all costs (due economic and fiscal concerns) but somehow was totally comfortable at sitting in inflation linked bonds. Equities are actually more resilient to inflation than inflation linked bonds
So emerging market and some developed market inflation linked bonds are not as safe as they look.
Out of the countries in my index, there is a lot more trust because those countries have more than 1 century of resonably stable inflation rates.
Last edited:
One way I could solve this would be to add a couple more currencies to the index, this would bring an equal weighted exposure to 12.5% of each. Precious metals would be 25% of the total, still quite high but closer to a more resonable level where the volatility starts to get dampened by the diversification effect. Potential additions are the Canadian Dollar (Canada its on the Rogoff never defaulted list) and the GBP (currency its been around a long-time, has inflation linked bonds and the defaults they did were minor)
Looking at the Purchasing Power Parities of the currencies I found
OECD data (vs USD)
Main currencies
CHF 140
DKK 120
AUD 118
USD 100
Additional (the "backup team")
CAD 94
GBP 99
So it appears that the main currencies command a premium over other currencies due their relative safety/stability. I know that has been the case for a really long time for the CHF.
I read about a investment guru whose PHD thesis was about what mattered in terms of dertemining the value of a currency. He found that two things mattered, PPP (Although I think he used the Big Mac index for this) and interest rates (I assume he used real interest rates).
He even designed a system that looks around for countries that have a good discount at the PPP level and high interest rates. Those currencies will tend to appreciate over others
But it's easy to see why, they are less stable and have a higher chance of wiping out the investor. Not that it isn't good to go for those returns, its ok, as a long its not with all of the capital (or a substantial percentage of the capital). That way risk of ruin is removed and one can try to extract the premium safely. That premium probably does exist but it isn't a free lunch, it comes with the higher tail risks (debt default, inflation/hyperinflation, other things)
Out the currencies in the index, its noticeable that they are pretty much all low yielders. They tend to be funding currencies for 'carry' trade strategies. They are the ones that tend to drop small amounts year after year against big yielders (like the BRL, ZAR) but they are also the ones that hold up in value when dislocation and instability appears.
Interesting strategy ideas arise when one considers mixing defensive FX strategies (focusing on those stable low yields) with aggressive asset strategies. For instance, lets say someone is long 50% stable high PPP low yielders and 50% the ZAR (South African Rand). With that 50% there is a high return strategy being implemented (say, long REITs).
Then the investor starts to get worried about the world and a potential crisis, but he is also has a high conviction that these REITs are too cheap and in a few years they will rise by 50%-100%. An interesting strategy is to, instead of selling the REITs, is to change the FX mix to 80% stable low yielders (through derivatives or FX pairs) and 20% ZAR. Shorting the ZAR through derivatives or FX pairs will be costly in terms of carry but that is more than made up by the high yield of the REITs. One can call that cost a 'insurance premium'. If the world does run into trouble, there will be a run to the stable low yielders currencies and the investor will have a FX gain (versus the ZAR and versus other currencies around the world, in other words, he is getting richer "in global terms")
That's an interesting way to cut down risk while maintaining that investment strategy.
I feel like currencies is the weakest part of my global macro game so I plan to explore more this topic and do some more research on this
OECD data (vs USD)
Main currencies
CHF 140
DKK 120
AUD 118
USD 100
Additional (the "backup team")
CAD 94
GBP 99
So it appears that the main currencies command a premium over other currencies due their relative safety/stability. I know that has been the case for a really long time for the CHF.
I read about a investment guru whose PHD thesis was about what mattered in terms of dertemining the value of a currency. He found that two things mattered, PPP (Although I think he used the Big Mac index for this) and interest rates (I assume he used real interest rates).
He even designed a system that looks around for countries that have a good discount at the PPP level and high interest rates. Those currencies will tend to appreciate over others
But it's easy to see why, they are less stable and have a higher chance of wiping out the investor. Not that it isn't good to go for those returns, its ok, as a long its not with all of the capital (or a substantial percentage of the capital). That way risk of ruin is removed and one can try to extract the premium safely. That premium probably does exist but it isn't a free lunch, it comes with the higher tail risks (debt default, inflation/hyperinflation, other things)
Out the currencies in the index, its noticeable that they are pretty much all low yielders. They tend to be funding currencies for 'carry' trade strategies. They are the ones that tend to drop small amounts year after year against big yielders (like the BRL, ZAR) but they are also the ones that hold up in value when dislocation and instability appears.
Interesting strategy ideas arise when one considers mixing defensive FX strategies (focusing on those stable low yields) with aggressive asset strategies. For instance, lets say someone is long 50% stable high PPP low yielders and 50% the ZAR (South African Rand). With that 50% there is a high return strategy being implemented (say, long REITs).
Then the investor starts to get worried about the world and a potential crisis, but he is also has a high conviction that these REITs are too cheap and in a few years they will rise by 50%-100%. An interesting strategy is to, instead of selling the REITs, is to change the FX mix to 80% stable low yielders (through derivatives or FX pairs) and 20% ZAR. Shorting the ZAR through derivatives or FX pairs will be costly in terms of carry but that is more than made up by the high yield of the REITs. One can call that cost a 'insurance premium'. If the world does run into trouble, there will be a run to the stable low yielders currencies and the investor will have a FX gain (versus the ZAR and versus other currencies around the world, in other words, he is getting richer "in global terms")
That's an interesting way to cut down risk while maintaining that investment strategy.
I feel like currencies is the weakest part of my global macro game so I plan to explore more this topic and do some more research on this