Register for free to join our community of investors and share your ideas. You will also get access to streaming quotes, interactive charts, trades, portfolio, live options flow and more tools.
Register for free to join our community of investors and share your ideas. You will also get access to streaming quotes, interactive charts, trades, portfolio, live options flow and more tools.
Thanks for the additional advice. I definitely agree that it's important to work through and understand the logic and implications of something first.
Many thanks for the very quick and unambiguous clarification. I'll think further.
LD AIM
The recent post from aaCharley has led me to have another look at the LD AIM spreadsheet. I gave up before, and am struggling again.
Looking at the Inputs side of the basic spreadsheet, I can't work out what to put in the Total $ Available box. Should it be the grand total of all my capital for investment, half of that to allow for the cash reserve, or the amount I have available for this specific investment if I was doing normal AIM rather than LD AIM, and if the last of these, should it be the whole amount including the 50% cash reserve, or just the 50% for the stock?
So, if I have $100,000 capital overall, and if in normal AIM I would be thinking of $10,000 to be allocated to this investment ($5,000 for the initial investment and $5,000 cash reserve), do I input $100,000, $50,000, $10,000 or $5,000 to the Total $ Available cell in the LD AIM spreadsheet.
Sorry to be dumb!
VWave getting high?
If the 90th percentile value for the VWave is 59.1 and this week's value is 55.98, I wonder what percentile that is - around 78? Is the risk level getting rather high?
Low Down A.I.M.
With Low Down A.I.M., is there any advice about how to decide how many buy and sell transactions to allow for? The numbers in the published sample spreadsheet are 4 sell and 3 buy. It seems to make a big difference to the $ Allocated and hence the Capital at Risk and Cash Reserve if you change these numbers.
If I've understood AIM correctly, with 10% buy SAFE and 10% sell SAFE AIM will not start to signal selling until current value has reached nearly 12% above starting value, but will start signalling buys after a fall of just over 9%, in both cases because the 10% SAFE is applied to current value not starting value or Portfolio Control. I hadn't realised this initially. Coupled with the additions to Portfolio Control after buys, this seems to give AIM a bias towards accumulation and portfolio growth - probably a good thing in an investment (rather than trading) system.
Other simpler (less subtle) core position trading and fixed value investing approaches seem to rebalance at 10% above or below the starting value and do not add any proportion of buys to the control value. The other difference is that they rebalance right back to the starting value rather than just to the edge of the Lichello band.
Nothing new here, I know, but I'm just checking my understanding.
I must admit I'm still quite partial to a variable ratio system based on 100 minus the percentile value of Shiller's PE10 within its range. Currently suggests 15-16% equities, since PE10 is in its 85th percentile. Whatever the actual level, trading the changes is interesting. I think of it as being a bit like a wave power generator. The surge generates power (throws off cash at higher prices), and as the wave recedes it sucks cash back in at lower prices. Since I am operating with an index tracking fund in an environment with no minimum trade size and no charges I can respond to relatively small changes, and will do so if it proves to be a profitable little generator.
I too have benefited from a large dose of government bonds in my portfolio during the last quarter of 2011 - just as I would have in 2008, if I'd had them then (I didn't!). That makes me wonder about a very basic 33% cash, 33% government bonds, 33% equities portfolio, rebalanced at 5% deviations. So simple and pretty safe, but not stellar.
There has been quite a lot of comment in the investmebnt press about the failure of gold mining stocks to follow the increase in gold prices, that the 'disconnect' is now way out of line with history and therefore a lovely catch up could be around the corner (and it seems gold miners have raised their gold price forecasts). BUT I read a comment that the disconnect could be due to the rise of gold ETFs - presumably because thay allow investors to invest in the gold price itself more easily, perhaps lowering the demand for gold mining stocks as an easily tradeable way, though indirect, of backing gold. If this is the case, the disconnect could reflect a new norm and the hoped for snapback may not happen.
Thanks for the responses to my previous post. I haven't responded but I have read and thought about the points made, and value the comments and advice.
Lack of SPY buy signal
Does this not mean that a standard 10% SAFE setting is not appropriate for all securities? Wasn't the default 10% set with individual stocks in mind, and I've seen a suggestion that a security with beta 1.4 is ideal for AIM.
A large part of the purpose of AIM, as I understand it, is 'volatility capture', so doesn't SAFE need to be set in relation to the volatilty of the instrument being traded? If SAFE is set too high, rather than "AIM being mean with its cash", isn't AIM as set up failing to capture potential profit?
If the beta of SPY is 1, perhaps SAFE should be set at 10x1/1.4=7. Some funds, e.g. bond funds could have even lower SAFEs on this basis, by relating the 12 month volatilities of the different funds.
Tom Veale wrote on the siliconinvestor BB on 6/10/2011, "Starting back at the beginning of the New Millennium I suggested in my old AIM Newsletter that people start shifting their SAFE emphasis from being biased for Accumulation to that of Distribution. Essentially I suggested that we drop the Sell SAFE to zero and shift all of it to the Buy side. Throughout the period since I've kept most of my AIM engines so tuned. So, if they were formerly 0% Buy and 20% Sell SAFE they shifted to 20% buy and zero percent Sell SAFE.
I wondered if we could automate this Split SAFE decision. Supposing a fund's price is 40% up the scale from a major bottom (early 2009 say) and 60% below the corresponding major top (late 2007 say) - I can give an example. If the combined buy and sell SAFE is 20%, you could make buy SAFE 0.4x20=8, and SELL safe 0.6x20=12, in other words biasing towards accumulation or distribution in a graduated way according to distance from major support and resistance. In late 2007 the fund would have been approaching the 1999 top and a long way from the spring 2003 bottom, so the same mechanism would have prescribed a low Sell SAFE and large Buy SAFE, because the risk of a fall looked much greater than of a continuing rise.
On the basis of this I have just reduced Sell SAFE on govt. bond funds I have to 0, and taken all profits, back to PC level, increasing Buy SAFE accordingly.
Regarding equities, I'm inclined to believe the view that the next secular bull market won't start until after the S&P 500's PE10 has dropped to single figures. That could imply a fall in the S&P 500 of about 60% some time in the next few years I think, so as an overriding safeguard I feel inclined to bias AIM towards that eventuality, in terms of both the cash proportion of new equity AIMs and the splitting of SAFE. Maybe this is too conservative?
Thanks for the various responses to my previous post.
Best wishes and Happy Christmas to all.
I have read about two other approaches which have some resemblances to AIM: Constant Value Investing and Core Position Trading.
In the description of Constant Value Investing that I read, an investment is allowed to rise or fall by 10%, then it is rebalanced back to the original cash value by either selling some and adding to the cash reserve or buying more from a cash reserve. AIM on the other hand (as I understand it) only rebalances back to the margin of the 10% band, and also increases the Portfolio Control when buying (by half the purchase value). When compared with Constant Value Investing, AIM would appear to (a) increase future buying because the Control figure is increased, but on the other hand (b) reduce buying because rebalancing is only back to the bottom of the 10% band, and (c) reduce selling by only rebalancing back to the top of the 10% band.
In the version of Core Position Trading that I read about, every time an investment falls by 10%, 10% more is bought, and every time it rises by 10%, 10% is sold, without reference to current and starting values. AIM and Constant Value Investing on the other hand do not sell until the original value (or increased Portfolio Control figure in the case of AIM) has been regained and exceeded by 10%.
I wonder if the is any theoretical and/or experimental work comparing the outcome of these approaches.
Appropriate SAFE levels
I've been thinking a bit about what are the appropriate SAFE levels for different types of mutual fund. I read somewhere that a good fund for AIM is one with a beta of 1.4, and original AIM uses 10% as the SAFE for both purchases and sales, so I'm using these two figures as baselines. Assuming a fund which tracks the broad equity market has a beta of 1, one could deduce a SAFE level for it of 10 x 1/1.4 = 7. Comparing the 1 year volatility for different funds with this general market fund, one could then derive SAFEs for them. The UK FTSE All Share Index fund I am basing this on currently has a published 1 year volatility of 17.1. UK government funds have published 1 year volatilities around 5, so one could derive SAFEs for such funds thus: 7 x 5/17 = 2. A gold mining stocks fund on the other hand has a 1 year volatility of 24.7, so the SAFE for it would be 7 x 24.7/17.1 = 10. And so on. I'd welcome comments.
AIM and technical indicators
Thanks for the very full response. One thing that attracts me to AIM is its simplicity and the fact that you don't need to look at charts and indicators if you don't want to, once you've chosen what to trade. Interpretation of charts and indicators usually ends up having a subjective component, which AIM needn't, and AIM keeps your attention focussed on your account value and profit and loss.
I've still got some UK government bond fund holdings and I decided to create AIM spreadsheets for them. I was quite surprised, pleasantly so, to see that in the last 4 months they have increased to the point where AIM with 10% sell SAFE is instructing some sales, which I've implemented. I had assumed that they probably wouldn't be volatile enough.
I've also started a Twinvest-based UK equities/cash account for my 7 month old grand-daughter. 18 or more years to run!
Cutler's RSI
Still migrating to an AIM set up, in a progressive rather than all at once way. Quite pleased (for the moment, in light of the bounce today) that I transferred some more into my equities fund over the last week. I used sub-30 readings on Cutler's version of the RSI (14 period) to prompt the purchases. Cutler's version uses simple moving averages (or just the sums) rather than exponential MAs and seems to give more/earlier signals. I added the whole value of the purchases to Portfolio Contol, as non-AIM events.
Tempting to consider selling some when the same indicator goes above 70 - a test I've done suggests it could be quite a profitable strategy! It wouldn't be AIM but it would have something in common with Low-Down AIM.
VIX and age
Thanks balbrec2. Your formula would have had me 44% in equities for September 2007, compared to 40% for V-Wave.
I have an alternative idea: 100 minus the Shiller PE10 percentile for the proportion in equities. For Sept 2007 this said 7% equities, but at the market bottom in March 2009 it said 65% equities. Your age plus VIX system said 68% - very similar, but the PE system had a much lower percentage during the large fall preceding the bottom, which would presumably mean a much smaller drawdown, but still providing the opportunity to put a lot in at the bottom. The PE system was still only 22% equities in October 2008 and went to 45% for November 2008.
I have to say though, that the PE system specified a very low % in equities from the mid 90s onwards, including only 14% at the other cyclical bear bottom in March 2003.
Best wishes.
Thanks Tom,
I completely agree about test periods. As well as being a short period, so much depends on the values at the start and end points of these comparisons. The one I chose starts right at the top and end of a cyclical bull, though it does include a cyclical bear and the recovery to date. When I've run the 10 month average against the FTSE 100 back to 1984 the results have been worse than buy and hold, though with less risk. Some things work in secular bear markets but not in secular bull markets (well you don't need anything then perhaps, but you can't be sure when they're starting or ending). Basic moving average systems do suffer in relatively calm sideways markets where you get whipsaws, as you say, which tends to happen as markets are turning. A number of small losses can add up to be significant.
I have found that a complex of moving averages of various lengths, looking at their relationships including convergence and divergence, giving a points score, applied to the S&P 500 on a daily basis, then using the signals to time trades on the FTSE 100, produces a return about 1.5x buy and hold over the whole period from 1984 to date, with less risk, on an all in, all out basis. Using the US index closing at 9pm UK time to time trades on the FTSE the next morning seems to give a slight edge. It can also be used for phased entry/exit. But having to check daily is a disadvantage. Weekly versions are quite good, but monthly is too slow I think. My benchmark for such systems is what did it do in October 1987!
I might make a separate post about a system based on Shiller's PE10 and the total market capitalisation to GDP ratio.
Hope everyone over there enjoyed Thanksgiving.
Best wishes.
Evaluation of AIM
I omitted the following additional point of comparison: the maximum drawdown on a month-end basis was 19.6% using AIM (in Feb 2009) and 5.5% with trend following (in June 2009).
Evaluation of AIM
Further to my earlier post, which recounted running standard AIM against the UK FTSE All Share Index from Sept 2007 to date, starting at 40% equities as suggested by VWave at the time, and producing a total return of 1.5%.
The ROCAR came out at 3.1, based on the average in equities of 51%. The arithmetic mean monthly return was 0.1%, with a standard deviation of 2.6.
I've just compared the above to using the 10 month simple moving average, as advocated by Mebane Faber, in a trend following system. If the index closed at the month end above the SMA10, 100% of capital was invested in the index for the following month. If it closed below the average, 100% was in cash for the following month.
For the same time period, ignoring dividends, interest on cash and trading costs, the total return was 30.2% and the ROCAR was 55, based on the fact that the system was 100% in equities for 55% of the months, and in cash for 45% of the months. The average monthly return was 0.6%, with a standard deviation of 3.0. Buy and hold was -13% as before.
Evaluation of AIM
I did a little test last night on using AIM with the FTSE All Share Index, starting at the end of August 2007, a few months after I retired with a lump sum to invest. The VWave cash level at that time was 60%, so I split the hypothetical account 40% equity index fund and 60% cash to start. I used the basic 10% for both buy and sell SAFE and used only month end figures and trades. The end result at the present time is that the combined equity and cash account is up 1.5% compared to the start, whereas a 100% equities buy and hold would be 12.8% down (a 40/60 buy and hold mix would be proportionately less down). I have not taken account of dividends, interest on cash or trading costs. At the worst point, Feb 2009, the total account value was 19.6% down.
A very unfortunate time to start of course. The Shiller PE10 was in its 93rd percentile, and the index was 17% below the top of its 2000/2003 range - both perhaps suggesting a higher cash allocation than VWave.
I might compare it to a an equities/bond mix and rebalancing regime to compare the outcomes.
Criticise it? No, I think it's interesting. Do you use a moving average of the VIX at all to smooth it out?
Hi Is7550,
Thanks for all your detailed points. I like the Permanent Portfolio too. With regard to UK government bonds, it's interesting to note that in 2008 they went up strongly, about the only asset class that did, and a portfolio about one third equities and two thirds gilts would have been very robust.
Best wishes,
Mike.
Hi Toofuzzy,
Thanks. They can all be taken as indicators for a starting cash:stocks ratio, like the VWave, and that's what I was suggesting, then follow AIM afterwards. Some of the links do also embody their own timing regimes, but these are based on similar principles to AIM - reduce stock holdings when prices increase and buy more stocks when prices and valuations fall. It might in fact be a simpler way of doing this than AIM.
Best wishes,
Daisy.
V-Wave alternatives
With apologies to those who know all of the following already, I can think of/suggest a few alternatives to the v-Wave for deciding the appropriate cash level:
1. Market Cap to GNP/GDP ratio (from Warren Buffet). See:
http://www.gurufocus.com/stock-market-valuations.php
and
http://www.myplaniq.com/LTISystem/advanced__indicators.action
with a suggested switching regime at:
http://www.myplaniq.com/LTISystem/jsp/strategy/View.action?ID=613
Currently 50%. I would prefer to use a more graduated approach than switching by 25% as soon as a threshold is crossed, but this strategy alone has reportedly produced very good returns over the last 1, 3 and 5 years.
2. Shiller’s PE10. See:
http://www.gurufocus.com/shiller-PE.php
and
http://www.multpl.com/
with a switching regime at:
http://www.myplaniq.com/LTISystem/jsp/strategy/View.action?ID=629
Again I would prefer a more graduated approach. Another possibility is to use the percentile value of Shiller P/Es to determine the cash proportion. This would give 81% cash at the moment, so very conservative.
Data on the average 10 year return from various Shiller P/Es could also be ranked and used:
http://www.mebanefaber.com/2010/06/09/shiller-pe-ratios-and-10yr-annualized-real-returns/
3. Another possibility is to used the ‘Scaled’ figure in: http://www.vectorgrader.com/strats/pmm.html
4. Finally, you can look at where the market is in relation to the various support and resistance levels. For the FTSE 100 this suggests 60% cash 40% stocks at present.
SAFE settings for mutual funds
Tom,
Thanks for your response, including the valuable additional point about leaving 30 days between purchases.
At the moment I'm progressively migrating my existing fund holdings over to my A.I.M. equity and cash funds, including from some government bond funds, trying to phase it so that I buy into the equity fund when the index is oversold, selling bond funds at this point too, when they are higher, until I reach the target cash figure, setting the Portfolio Control figure from the value of the equity fund holdings at that point.
Best wishes, Mike.
SAFE for mutual funds.
I am a retired, UK-based investor who is just starting to use AIM. I'm thinking of sticking to a FTSE All-Share Index tracking fund for the moment. It is relatively cheap, has no switching charges in the environment in which I'm buying and selling it, and is highly correlated with devloped world stockmarkets. I anticipate diversifying into, for example, an emerging markets index tracking fund if things seem to be going OK.
In a post to the Advanced users BB in August 2001, Tom Veale stated that he used a total SAFE (buy plus sell) of 10% for funds which follow the broad market.
I wonder if he would still recommend that. It had already occurred to me that relating the SAFE parameters to the volatilty of the security being traded seemed to make sense.
Inflation indexing of Portfolio Control
Any views about increasing Portfolio Control to keep pace with consumer price inflation?
See http://beand-cpt.blogspot.com/2007/12/inflation-indexing-of-aim-portfolio.html