Cash is Risky

You can’t get away from risk unless you are dead, which is fairly high price to pay for eliminating risk.  A portfolio of cash just has different risk exposures than a portfolio of other assets.  Cash as a component of a portfolio makes all the sense  in the world.  It increases robustness and allows you to take advantage of situations as they arise but if you think because you hold a boat load of cash (or deposits or t-bills) that you are risk free; you are wrong.  Consider the following picture, it shows the real value (inflation adjusted) draw-down (or value relative to its previous peak) of a T-bill portfolio.

What does this mean?  We’ll for most of our lives T-bill have been fairly safe investments but in the early 1940s they were terrible and although you didn’t lose the nominal (number value) of the investment you did lose it purchasing power as inflation ate away at what you could buy. You can see this illustrated by the next chart which shows year over year inflation and T-bill rates.

Maybe you now believe me that cash is risky and I have not even illustrated the real risky case studies for fiat cash holding: Currency collapses.  Think Germany after WWI or Argentina a few years ago or Russia or (this is a long list).  Investors who held thier wealth exclusively in cash were devistated. 

Thus I repeat my intro, you can’t get away from risk, you can only manage it.  Sensible risk based diversification is the means to do this.

Is the stock market crash good for you?

Having less wealth because the assets you own have dropped in value (specifically real purchasing power) is on its face bad, but what if the change in asset price valuations are not a reflection of a change in long run economic perceptions about growth but only a shift in the price of risk.  If the returns for taking risk increase the prices of risk assets must drop in order to bring up the yield.  If this is the reason for the stock market crash (it is in part but only part of the reason) then certain people actually benefit from the stock market crash.  Lets say you are young (or uber wealthy) and you really only care about the value of your retirement account in 30 years. (Think about who this would be…wealth transfer from older savers to younger savers).  If you were given the option of having your saving growth at a 4% rate or cut in half but grow at an 8% rate (the analogous situation to stock market crash) which would you take?  You should take the hit up front and the higher return.  Now life is never this simple but it illustrates a point. 

 

The X-Axis is in years from the start date.

Front Running and Policy Response

Each January someone writes about the January Effect or the reality that stocks tend to rally the first week of January.  A number of influences are underlying this result such as tax planning (the tendency to sell your losses – to realize the tax benefit in the past year – at the end of the year, this pushes additional equity supply on the market during the last week of December and when that selling pressure drops off the stock may rally).  Another plausible story is that bonuses come out at the end of the year and to the extent that people save a lot more of their bonus than their normal salary (and they saving some of it in equities) then a larger inflow of cash into equities during first week of year may be expected (seems like an upward pressure).  Now each year the stories focus on these narrow explanations and seem to wonder at the feasible underpinning of the efficient market hypothesis (EMH), but then fails to extend the analysis.  Think about it for a moment…What other places do flows result in risk arbitrage opportunities (and if a well recognized one works, what about a poorly recognized one)?  The lack of speculation or interest when these stories are written always surprise me but that not where I’d like to go.  I’d like to consider the broader consequence of non economic flows that result from policy decisions.

What does it mean that capital allocators are simply trying to front run (buy before others buy or sell before others sell) structural inefficiencies in that market place?  What does it mean that prices move based on non-economic factors?  Should policy makers (the rule setters) care?  I think so and here is why.  Capital allocation (profit seeking in the financial work) is valued added when it is providing risk arbitrage and bring expertise to capital deployment; it is less beneficial to the system when capital is used to smooth out structural or transactional inefficiencies; first because it takes capital away from real economy activities and second because when everyone tries to front run the result can become arbitrary price movements which will cause both poor capital allocation and increase systemic risk (to the extent the new randomness is not percieved).  Note the non economic flows do provide liquidity which is valuable. 

What can policy do?  Well it could try to encourage robust pools of capital through tax or regulatory policy.  For example a Warren Buffett whose uses of capital is a good example of what I mean by robust pool of capital.  These pools could be given a lower capital gains tax rate than say a momentum based trader, which would encourage capital to move toward those economic strategies.  Another possible policy would be to minimize the discrete breaks in tax policy by eliminating what is the first week of the tax year.  Consider if the tax year was assigned randomly then there would be no standard January period.  Other policy options include better regulatory data management.  All of these are tricky to implement and may not be worth the cost but I think it is a question worth asking.

Ten Years of Uncertainty

Today’s FOMC releasediscussed the Fed’s decision to continue the quantitative and qualitative easing programs which have allowed the Fed to play the role of the banking system-performing credit and maturity transformations-while ensuring that the money supply doesn’t collapse.  The release also mentioned that the Committee was “evaluating the potential benefits of purchasing longer-term Treasury securities.”  That may have triggered the large rally in treasuries that pushed the 10-year yield down to 2.27% (down 25 bps). With this in mind, I’d like to expand on yesterday’s post.

First, consider that an arbitrage relationship exists between a 10-year rate and 40 sequential 3-month rates.  You could finance the purchase of a 10-year bond with funding from the next 40 3-month rates.  Next, imagine that market expectations are that the 3-month rate will be 25 bps (the high end of the new fed funds target range) over the next 10 years.  If the market were certain, then the ten-year rate would be 25 bps as well, but the situation changes when those expectations are uncertain.  Let’s say that market expectations are normally distributed with annual volatility of N (a stylized assumption to illustrate a point).  As the range of plausible 3-mth rates expands, the bottom of the distribution is cut off by a certainty, the zero nominal rate.  The market 10-year rate ends up being a value based on the weighted average of the expected rates. The truncated distribution thus moves the expected rate away from the first guess (25bps).  This can be thought of just like an option (selling the bond has limited downside in a nominal sense).

You can now see how uncertainty affects rates.  How much does this upward pressure on rates matter? The soft floor on the 10-year rate will be a function of the uncertainty about the future path of the 3-mth rate, as illustrated by the following chart.  The open question is what you think the proper degree of uncertainty is for the 3-mth rates.

 

Update: It is worth nothing that Japanese 10-yr rates fell below 1% (although they’re usually in the high 1% range).  One plausible explanation for this is that in a deflationary environment (inflation is running at -5%) where the nominal short rate is 0% but in real terms it is higher at 5%, it is hard to imagine the monetary authority actually wanting a tight monetary policy.  This may give you a large degree of confidence that the nominal rate will not increase, meaning that the distribution contracts and thus the 10-year rate is able to fall.

Federal Reserve Challenge: When normal just doesn’t cut it.

When the economy slows, the first responder is the Federal Reserve. It eases monetary policy by lowering the bank reserve rates (it is often tight monetary policy which triggers the slowdown in the first place).  This time around, the Federal Reserve responded more aggressively than usual, but they quickly realized that in this cycle problems would not be solved by decreasing the interest rate paid on loaned reserves.  As the reserve rate was cut, the rates that were paid by consumers and businesses did not drop-in some cases they even increased because spreads increased faster than rates fell.  The Fed’s efforts were ineffective.  Now the rate is approaching the 0% nominal floor, which means that this type of intervention (in a nominal sense) is about to run out of bullets. 

 But the Fed did not settle for the standard policy tools; it has gone well beyond what is normally done to stimulate the economy.  One means to stimulate credit is to guarantee bankers access to liquidity, so bankers need not worry about liquidity risk when doing maturity transformations (borrowing short and lending long).  With this in mind, and a focus on preventing widespread banking failures, the Fed started to push out liquidity to banks through programs such as the Term Auction Credit program (after discovering that no one would use the discount window).  This program was useful in that it held the system together (we would be a lot worse off without it), but for the past few months the Fed hasn’t been able to push money out the door. This means that for every 100 dollars it auctions, only a fraction, say 60 dollars, is bid on.  Banks are so focused on deleveraging that they have no interest in doing maturity transformations even with cheap Fed money. 

 As things continued to fall apart, certain markets, such as commercial paper, were on the verge of collapse. The Fed’s response was to set up a program to provide liquidity so that bankers could buy the CP, but it quickly realized that it wouldn’t work because bankers were not interested in expanding their balance sheets.  The Fed then set up a program to take the CP on its own balance sheet.  We have now reached the stage where the Fed has taken on the maturity and credit transformation responsibilities which were once the role of the banking system.  Examples of this include the money market investor facility, the commercial paper facilities, and the new GSE purchase facility (and possibly a program to buy long term government debt).  In addition, the Fed has held together the fragile banking system by facilitating the JPM-Bear Stearns deal and providing ‘liquidity’ to AIG. 

 Now that I’ve gone over what the Fed has done so far (and how unconventional its actions have been), a fair question would be ‘what else could the Fed possibly do during its meeting on Dec 15/16?’  It could expand the existing programs and commit to buying more risky assets (thereby committing to do more maturity and credit transformation).  Even though this seems like more of the same, it would be beneficial.  I think the Fed should also provide transparency to the market on the duration of these programs.  While this would require the Fed to keep these programs in place or interest rates low for a certain period of time, which bears some risk, it would also allow the market to start pricing in the continuation of these programs and possibly encourage bankers to take advantage of the subsidies provided.

 To show the effect that future assurances could have, look at the forward interest rate picture.  The most likely path as imputed from Fed Futures (financial instruments used to transact on future rates) is that rates will fall to 25 bps (0.25%) and stay there throughout 2009.  But the futures curve for the same interest rates has them priced at around 80bps (0.80%) at the end of next year.  Why the difference?  Because as time increases, so does the uncertainty surrounding the expected rate.  If you look at the -1 SD (standard deviation) and +1 SD line, you’ll see a plausible distribution for future rates.  Think about it this way as the prediction is about a further point in the future you become increasingly uncertain about the prediction.  The expected rate ends up being the average of this distribution of rates. The problem is that the low side is truncated at zero so even if one does not expect rates to increase, the fact that your uncertainty increases as the duration of your prediction increases means that the forward curve prices in higher rates.  If the promise were made to hold down the rate at 25bps through 2009, then the forward curve would drop because the +1 SD deviations would drop (based on our confidence in the Fed’s statement).

 

Remember flatting out the short rate curve will pull down long rates.

Snapback: The End of Deleveraging?

The problems in the financial sector that started in mid-2007 have created problems for the real economy, and the problems in the real economy are creating more problems for the financial economy. This self-perpetuating downward spiral is in full force.  Along with consistently negative surprises in economic data, this state of affairs has many commentators saying, ‘it’s going to get worse before it gets better.’ I find it annoying that they omit to specify which part of the economy they are talking about. Is the real economy going to get worse before it gets better, or is the financial economy going to get worse before it gets better? If they are referring to the real economy and suggesting that GDP is going to fall in real terms or unemployment is going to rise, then they are really not worth their pay. One look at the economic data will give you the short term direction of the real economy.  If they are making a prediction about the financial economy, then I’d be curious to understand their logic. On what basis do they think that asset prices will continue to fall? And how can anyone confidently make this claim?

One coherent argument is that the deleveraging cycle is not over, and that continued equity write-downs along with funding withdrawals will create more and more forced selling.  As long as expectations about deleveraging are an overhang on the market, there is not much incentive to buy because you expect to get a better price down the road.  Forced selling does two things to the market: (1) it creates an inelastic seller who will push down the price, and (2) it creates an environment that pushes away buyers.  If you know that someone has to sell a large chunk of stock, you’d rather buy the last piece of it than the first.  You will only buy early if you are given a steep discount from the present price.

This analysis is sensible enough, but because we are well along into the deleveraging process I don’t think it validates a short position-or at least not a significant one-in risky assets.  At some point the forced selling will stop, and when it does I expect the upward volatility to be higher than it typically is.  The risk for short sellers trying to take advantage of the deleveraging in the market may be higher than they realize; not only does forced selling push the price down, it also scares away buyers.  When the deleveraging is over, not only is the inelastic seller gone, but those buyers who had retreated will come back to the table. This is also the reason why downward volatility has been so nasty.

In an attempt to demonstrate this phenomenon I’ve compared the upward and downward volatility over time.  Admittedly there are some conceptual problems doing it this way, but the graph is useful.  Look at the bottoms in the stock markets, rallies off the bottom are more violent than other upward movements, meaning that the downside minus upside volatility (the blue line) is negative around a bottom (1987 being a major exception).  The red line is the S&P 500 in log space.

Side Note 1: One more reason forced selling creates a terrible market: A wise friend was waiting on the side lines with a large pile of cash, thinking that the Dow would fall to 10,000 in this downturn and likely would have considered investing around that point. But the Dow fell from 11,000 to 8,500 in the span of a week. The idea of trading on valuation was offset by an increased sense of the unknown. Although the 10,000 mark seemed like a good value when he was watching the markets in early 2008 (when the Dow was at 13,000), his view on his own understanding of what was going on and thus his willingness to take risks changed when it fell to 8,500.  When things don’t make sense you should step back. As the adage goes, invest in what you know.

I AM LEGEND

Money managers love to quote statistics suggesting that their performance is not luck but skill.  Many of them acknowledge that randomness does have something to do with performance and with a enough people someone who are not will appear skilled (the classic quarter flipping example).  But after telling this story they use the same statistical framework to show how they are really talented.  Usually either by showing that they have beaten the index for X years or having returned excess alpha for Y years.  I do not believe that these fellows are being disingenuous and the methodology isn’t crazy but I would argue that one should look closely at the investment process so you understand their advantage before evaluating the statistics.  One reason for this is described below. 

Imagine someone who knows that they are not talented – however unlikely that is – but they wish to be a legendary money manager.  This individual sets about to find a strategy that will give them the best chance of appearing to have an alpha CAGR (compound annual growth rate) of 10% for 28 years, which on top of a risky beta would likely give a low teens real return.  Doing this over 28 years would likely get you some attention.   (Note: It would not be real alpha but just have the appearance of alpha). 

Other managers would bring out their statistics arguing that given no talent creating any alpha return for each year over 28 years is nearly impossible (1 in 268 million).  Well that’s not true a simple martingale strategy could easily create the trivial appearance of beating a benchmark and work for well more than 28 years, particularly if applied on a daily basis.  The managers would likely come back and argue that creating 10% alpha over 28 years is nearly impossible and you can’t hide behind a trivial beating of the benchmark when trying to create 10% annual excess returns.  Our reflective no talent dreamer would come back at them and admit that it is unlikely but an optimal strategy will give him a much better chance than they are implying.  What is this optimal strategy: bet big.  If he takes enough leverage he can give himself a chance.  Admittedly the expected returns (the purple line – geo average (R-axis)) because he will be well past the risk of ruin will be bad but the chance he hit my goal will be optimized.  The chart illustrates the reality based a simulated random strategy on S&P 500 excess returns.  (Left Axis being the probability of being above some level of alpha returns, given some level of leverage (x-axis), while the purple line shows the expected return from this random strategy). 

I’ll add one more chart to help conceptualize this reality.  This chart is based on an investment strategy with an Sharpe ratio of 1 at various degrees of leverage (the x-axis, 6 being 600% leverage).  Each line is the simulated standard deviation meaning where the first line (purple) is where 2% of returns would be over that, the third line (dark blue) is the mean expected value, and the fifth line (red) is where 98% of returns would be above that point.   The Red dot is the geometric mean max point, which means if your goal is to have 50% excess returns you should past the geometric mean maximizing leverage.  (I think this red point could be a key to financial regulation but that is another story).

 

So if you are a no talent trader and you want to win big, take ridiculous large – negative expected value - bets, but even if you are talented (a true alpha Sharpe ratio of 1 is talent) and your desire of stardom is big enough pursuing the same strategy make sense.  Remember this reality the next time your star trader tells you the key to his success is having the confidence to bet big.

Lotteries: The Price of Hope or a Tax on the Poor

The State Lotteries at the Turn of the Century: Report to the National Gambling Impact Study Commission presents data on the heaviest players.  It highlights the following divergences between the distribution of the heavy players and their proportions of the population:

Characteristics of the top 20% of lottery purchasers, 1999
Demographic Group	Percentage of Heaviest Players	Percentage of US adults
Male	61.40%	48.50%
Black	25.40%	12.20%
High School Dropouts	20.30%	12.30%
Household income under 10,000	9.70%	5.00%
Median Age	47.5	43

 

Typically the focus of this presentation is on the low income line and the argument follows: lotteries are a regressive tax on the poor, albeit a voluntary one, which transfers money from the users of the lottery (implying foolish) to the wealthier (in forms of education subsides and other public programs).  People argue that on this basis that they should be banned.  This crossed me the wrong way in that I don’t like the government telling adults what they should or should not be able to do.  (of course the question of what an adult is, is in and of itself a challenge)  But what if it is not foolishness but rather the hopeless buying hope. 

What I mean is that I often hear people say, ‘I would never buy a lottery ticket, it has negative net present value.  How could someone be so stupid?’  We’ll its not just about the present value, it is also about the distribution.  In a portfolio you would take on assets even if they have negative expected value if they have really good correlation characteristics as it would improve the portfolio.  In non-finance term these groups are buying lottery ticket not because they think they are a good value but because they give them hope; hope is worth something. 

Now if you think I’m being paid by the New York State lottery – I’m not – let me clarify.  Let’s say I’m a high school drop out (just didn’t have the inherent mental capacity) who works a job that pays 10 dollar an hour, I work 50 hours a week and 48 weeks a year (so I work pretty hard); let say I’m paying total of 20% of income in taxes, leaving me with 1600 a month in income; I live pretty light (600 rent, 200 insurance, 300 auto, 200 food, 200 other/emergencies); meaning I could save up to 100 dollar a month.  Now if I save that money and invest it by the time I’m 65 (assuming this fellow is 25, and a real rate of return, after tax, of 3%) he will have just under 100K (in today’s dollars).  Not exactly a tremendous sum, meaning that he has no hope of financial security (or the high life) even though he’s making good decisions given his constraints.  So if he spends 52 dollars a week on lottery tickets he improves his chances from just about 0% to just about 0% (but he can identify the source of that improvement and believe in it) [note: the value of his portfolio at age 65 drops by about 4,000 dollars].  If he’s rational he doesn’t expect to hit it big but he now feels like (in large part because he can identify the source of the trivial probability) he has a chance.  (I’m not suggesting this is the best mean to give yourself hope but it is a means). 

So look at the table up top again.  Are these groups associated with foolishness or hopelessness (or something else, it is quite possibly driven by other factors)?  It is some combination of the two on average but at an individual level its likely one or the other (if instead of spending 1 dollar a week he drops half the paycheck on the table; it is foolishness).  So another argument, besides the libertarian argument against banning lottery sales, the fact that banning them would remove a source of hope for the hopeless.