In a typical deal a sponsoring bank moves loans off its balance sheet and sells them to a special-purpose entity. For legal and tax reasons this is a non-profit vehicle; if it is established under US law it is usually known as a ‘trust’. The special-purpose entity then sells bonds to investors, known as ‘tranches’ (from the French word ‘tranche’ – ‘slice’) of different seniority. The most senior tranches have the first claims on the interest and principal payments from the loans. The proceeds from the bond sales are then used to purchase the loans from the bank.
The senior tranches of these securitizations are extremely safe. They are protected in several ways: first, by ‘over-collateralization’ – that is, the practice of putting a larger value of loans into the special-purpose entity than the value of bonds sold – second, by the interest margin earned because the interest rate paid on the loans is always higher than the interest rate paid on the tranches; and, third, because of the seniority of the senior bonds. In a typical structure the most senior AAA tranches are only about 75 per cent of the total issue. This means that they are virtually free of any default risk; the losses on the loans would have to eat up all the interest surplus, all the over-collateralization and the 25 per cent of more junior tranches before the most senior tranches lose money. Loan losses of this magnitude would be quite extraordinary.
Sponsoring banks often keep the riskier tranches of securitizations themselves. This is a reasonable thing for them to do if the main motive for creating the structure is, as it usually is, obtaining cheap funding rather than transferring risk. They may also buy and sell loans in and out of the mortgage or other loan pool, in order to maintain the asset quality. There is often no legal obligation for them to buy bad loans and replace with good loans, but doing this helps to maintain a reputation for quality and thus helps when selling loan-backed securities in the future.
Even if the sponsoring bank retains the riskier tranches it still benefits from the securitization, because it can replace relatively expensive wholesale borrowing on its balance sheet with the relatively cheap funding from selling senior structured securities. The difference in cost can be very large; a bank might pay a spread of, say, 25 basis points above the standard London Interbank Offer Rate (Libor) on the senior tranches of a loan securitization but a spread of 150 basis points or more on floating-rate bonds issued on its own balance sheet.
Commercial banks and the structuring departments of the investment banks created a lot of paper during the boom years of structured and mortgage-backed credit, from 2002 until 2007. While there are no comprehensive statistics, industry bodies have recently released estimates of the outstanding stock of most categories of these securities. With some further estimation of remaining stock, based on the issue flows, it is possible to get approximate figures for the overall size of the market.
There are many acronyms in the world of structured credit.
The largest market is for so-called ‘agency RMBS’ in the United States – the residential mortgage-backed securities issued by the government sponsored enterprises Fannie Mae and Freddie Mac. $5.9 trillion is a huge number. It is about half of all US mortgage lending and not far shy of half of US national income (which was close to $14 trillion in 2007). But these agency RMBS are a special case; the mortgages that back them are guaranteed by Fannie and Freddie, so they are backed at least implicitly by the US government.
The next biggest market, especially in the United States, is the $2 trillion market for asset-backed securities (ABS). These are securitizations of retail lending of all kinds including vehicle loans, credit card receivables, student loans, equipment leases, small business loans and also home equity loans.
There are some other surprises from looking at these numbers. The sub-prime residential mortgage-backed securities, where all the credit market problems first appeared, are only around one-tenth of the entire market for mortgage-backed and structured securities, less than half of the ABS market (sub-prime RMBS are issued only in the United States, not in Europe) and less than the market for commercial mortgage-backed securities.
The broadest category here comprises ‘collateralized debt obligations’ or CDOs. They include a range of structures where the securitized assets are neither retail loans nor mortgages, for example restructured securities (such as the ABS-CDOs), corporate bonds, corporate loans and so-called ‘synthetic CDOs’, where the assets are the tradable insurance contracts known as credit default swaps.
Even if we exclude the agency-backed RMBS, there was close to $7 trillion outstanding, which is about twice the size of the market for tradable US government debt, normally regarded as the biggest and most liquid securities market in the world.
Bank exposure to structured credit markets
Basis spreads
The spreads between benchmark rates such as LIBOR and Treasuries vary over time. This creates opportunities to bet on whether these spreads will widen or narrow. The most practical way to trade on US$ rates is by using Eurodollar versus US Treasuries interest rate futures. If the spread between LIBOR and US Treasuries is expected to widen buy Eurodollar interest rate futures and sell Treasury futures.
Interest rate direction
If a trader expects the general level of interest rates to rise there are a number of ways to take a position on this. These include buying an interest rate swap where the trader is a fixed rate payer and floating rate receiver, buying FRAs, shorting liquid long-duration risk-free bonds and selling interest rate futures.
The Study of Trends
At its heart, technical analysis represents the study of price trends (or anticipated trends). In price terms, at their most basic, these can be divided into uptrends and downtrends. Within such trends, we see points where little price action occurs and conversely other points reflecting substantial price action and market tension. The idea behind support and resistance is that if the price action fails to exceed a certain level, then that level becomes important. Thus, if a price fails to exceed a high and falls back, we call that high a resistance. Equally, if the price action fails to get below a low price level, then that low price level becomes support. Price trends reflecting a number of support and resistance levels are reflected by trend-lines.
Resistance or support can be formed around such a trend-line. From this, we can say that it has broken trend-line resistance. Thus, we can describe support and resistance levels as levels where a trend may be interrupted or reversed. Because such levels can determine the continuation or the cessation of a trend, they are seen as important by market participants. In this example, market participants may well have left stop loss orders to buy Euro and sell dollars above the trend-line resistance on the view that if such a level broke it would signal a short-term end to the downtrend. Of course, if enough people leave orders to buy (sell) above (below) trend-line resistance (support), then the reversal of the previous trend could well be accelerated. Furthermore, speculative elements could discover such orders and try to target them in order to cause what might become a self-fulfilling move, allowing for potential profits.
To identify support and resistance levels, technical analysts use a variety of information inputs, including but not exclusive to chart analysis and numerical rules based on previous price performance. The rule with support and resistance is that they are important until they are broken. This may seem like just stating the obvious, but the key thing to note is that there is no particular time limit to their importance.
Currency Order Dynamics and Technical Levels
Sceptics may suggest that support or resistance levels can just as easily be randomly picked. The evidence however does not support such scepticism. Indeed, on the contrary, both academic and institutional research suggests exchange rate trends are interrupted or reversed at published support and resistance levels much more frequently than is the case at randomly picked levels. Such levels are therefore seen as statistically important, most likely because of the clustering effect mentioned earlier. Customer orders are placed just above or just below previous highs or lows. As a result, this clustering can have the effect either of pausing or accelerating the short-term price trend at any one time. This link between capital flows and technical chart levels can be expressed in the following way:
“Support” reflects a concentration of demand sufficient to pause the prevailing trend
“Resistance” equally reflects a similar concentration of supply
However, this clustering effect on prices can be further broken down into two specific types of customer order:
Take profit
Stop loss
There are important differences in the way that these two specific types of customer order tend to cluster. For instance, take profit orders tend to cluster in front of important support or resistance levels and thus tend to have the habit of causing the trend to reverse — thus reinforcing that support or resistance — if they are sufficient in number. By contrast, stop loss orders tend to be clustered behind important support or resistance levels, thus accelerating and intensifying the prevailing trend if triggered. Academic research has found that take profit and stop loss customer orders, which impose some degree of conditionality on the order, can make up between 10 and 15% of total order flow. As a result, they can have an important effect on trading conditions and therefore on price patterns. During calm market conditions, they can further restrict price action. Conversely, during volatile market conditions, they can exacerbate price volatility when such orders are triggered. Thus, both in calm and volatile market conditions, they re-emphasize the original importance of the support and resistance levels.
So far, we have been looking at “spot” foreign exchange orders, that is conditional customer orders to be executed for spot (T + 2) delivery. However, conditional orders left in the options market can also impact spot currency price action. More specifically, “knock-in” and “knock-out” levels for exotic options, allowing a client to be knocked-in to the underlying structure or conversely knocked-out of it, can and do trigger specific spot currency price activity. Knock-in and knock-out levels are usually chosen based on previously important highs and lows. In other words, they are chosen based on technical support or resistance levels. As a result, there can be — and frequently is — both spot and option customer order clustering around such levels, further impacting price action.
It is not only customers that place conditional orders in the market. In order to limit a bank’s balance sheet exposure to overnight price swings in exchange rates, interbank dealers either close out their positions at the end of the day or alternatively themselves leave take profit or stop loss orders with their dealing counterparts within the bank in the next time zone. Thus, a dealer in Singapore may pass on their customers’ conditional orders as well as their own to London and London may in turn pass on such orders to New York and so on round the time zones, either until such orders are filled or conversely are cancelled. If there is a self-fulfilling aspect to this whole idea, it concerns therefore the very microstructure of the currency market itself. Broadly speaking, currency interbank dealers follow technical analysis more closely than the customer base of the bank, in part because they have a much shorter time frame than their customers and in part because they have to trade in order to make a living irrespective of whether or not there have been changes in economic fundamentals. Currency interbank dealers and short-term traders follow technical analysis, and because they make up the majority of currency market participants the levels and types of analysis that they follow automatically become important. Thus, structural aspects within the currency market may help explain to some degree the success of technical analysis. What it does not explain however is the superior degree of that success relative to classic economic analysis or alternatively to random walk theory in predicting short-term exchange rate moves. Given that take profit orders cause price trends to pause, while stop loss orders extend such trends, the logical conclusion is that the balance between such orders in the market place is an important real-time determinant of exchange rates.
THE CHALLENGE OF TECHNICAL ANALYSIS
Technical analysis has posed a challenge to economic analysis in its ability to predict exchange rates. As a result, considerable research has been undertaken by the economic community on how technical analysis works, both in practice and in theory. It is not for here to go through this research or literature in detail. Rather, we look at one such study as symptomatic of a general enquiry by the economics profession into the workings of technical analysis. More specifically, no less than the Federal Reserve undertook to examine this phenomenon, apparent confirmation of an ongoing change in the way both private and public institutions are approaching the field of technical analysis. Indeed, the reader can find no more useful and detailed investigation of the subject matter, starting from a macroeconomic perspective, than two reports by Carol L. Osler of the Federal Reserve Bank of New York, which examine how technical analysis is able to predict exchange rates. These papers go a substantial way in explaining how technical analysis works and are particularly useful as they undertake this investigation from an economic perspective. In line with work done on studying order flow they suggest customer orders “cluster” around certain price levels and that such “clustering” creates specific price patterns depending on whether or not those levels hold. To a technician, this makes perfect sense given that a price represents the consensus of market supply and demand at any one time. Below the price, there should be “support” levels at which demand is expected to exceed supply and conversely above the price there may be “resistance” levels, where supply may exceed demand. From my perspective, I would suggest the following reasons why technical analysis has gradually taken on a more prominent and important role in predicting exchange rates:
Over the short term, the currency market is essentially trend-following.
The majority of market participants are speculative, that is they undertake currency transactions that have no underlying trade or investment transaction behind them.
Nature abhors a vacuum — currency market participants have to trade off something whether or not there has been any change in macroeconomic fundamentals.
Traditional exchange rate models have had relatively poor results, therefore another analytical discipline was needed that was able to achieve better results.
Exchange rate supply and demand create price patterns, which in the absence of other stimulus may provide clues for future exchange rate moves.
There does appear to be a crucial self-fulfilling aspect to technical analysis, which is to say that because a large number of people see a particular price level as important, therefore de facto it becomes important. Needless to say, this is an aspect that critics of technical analysis regularly seize on. While this may be the case to an extent, it does not answer the obvious question of why such a number of people find those levels important in the first place. Technical analysis is the discovery of patterns within price action, patterns which can be used to predict future prices. The predictive results of technical analysis consistently exceed those suggested by a random walk theory.2 Indeed, such have been the results achieved that there is now a sizeable and ever growing community of traders and leveraged funds that trade solely on the back of technical analysis signals. In short, technical analysis “works” to the extent that it produces results consistently for market participants who are trying to predict short-term exchange rate moves. If this is the case, what precisely is technical analysis and how can one use it?
ORIGINS AND BASIC CONCEPTS
At least in its modern version, technical analysis is generally seen as emanating from the “Dow Theory” established by Charles Dow at the start of the twentieth century. The core original ideas of technical analysis focused on the trending nature of prices, the idea of support and resistance and the concept of volume mirroring changes in price. Though we only touch on it here, the contribution of Charles Dow to modern-day technical analysis should not be underestimated. His focus on the basics of security price movement helped to give rise to a completely new method of analysing financial markets in general.
The basic premise behind this is that the price of a security represents a consensus. At the individual level, it is the price at which one person is willing to buy and another to sell. At the market level, it is the price at which the sum of market participants is willing to transact. The willingness to buy or sell depends on the price expectations of individual market participants. Because human expectations are relatively unpredictable, so the same must be said for their price expectations. If we were all totally logical and could separate our emotions from our investment decisions, one should assume that classic fundamental analysis would be a better predictor of future prices than it currently is. Prices would only reflect fundamental valuations. The fact that this is not the case suggests that other forces may be at work. Indeed, investor expectations also play a part, both at the individual level and also as a group.
Technical analysis is the process of analysing a currency or financial security’s historical price in an attempt to determine its future price direction. It is founded in the belief that there are consistent patterns within price action, which in turn have predictable results in terms of future price action. In contrast to economics, technical analysis requires that financial markets are not perfectly efficient, that there is no such thing as perfect knowledge or perfect information availability or usage, and also that in the absence of other information market participants will look to past price action as a determinant of future prices. For precisely this reason, the economics profession generally has dismissed technical analysis as irrational. However, just as we have already seen that financial markets are not perfectly efficient, so substantial research has shown conclusively both that technical analysis is widely practiced by market participants and perhaps more importantly that it has yielded substantially positive results. Traders who have used technical analysis have frequently made consistently high excess returns. Furthermore, in the context of the currency markets, technical analysis has a particularly good track record in predicting short-term exchange rate moves. How can this be so? Simply put, nature abhors a vacuum and thus in the vacuum left by classic economic analysis, in its inability to predict exchange rates over the short term, came technical analysis.
SPECULATIVE AND NON-SPECULATIVE FLOWS
While these flow and sentiment models vary, both in terms of the time span they focus on and the kind of information they look at, the basic premise behind them is the same — exchange rates are determined by the supply and demand for currencies, in other words by “order flow”. Over time, economic fundamentals will dictate the order flow and therefore the exchange rate itself. However, currency market practitioners do not necessarily have that long to wait. Therefore, it is necessary to study order flow separately and independently from the fundamentals, and moreover it is necessary to study the drivers of that order flow.
The key distinction between a speculative and a non-speculative capital flow, keeping to the definition that we are using for speculation — which is that speculation involves the buying and selling of currencies with no underlying attached asset — is the exchange rate itself. For a speculator, the exchange rate is the primary incentive for investing, using this definition. However, for an asset manager, the exchange rate is not the primary consideration, which is the total return available in the local markets. As the barriers to capital have broken down and as currencies have been de-pegged and allowed to float freely, so both speculative and non-speculative capital flows have grown exponentially. There remains a dynamic tension between the two, allowing one or other to be more important in terms of total flows at any one time.
Generalizing somewhat, one can say that speculative flows dominate short-term exchange rate moves, while non-speculative flows that are attracted by long-term fundamental shifts in the economy dominate long-term exchange rate moves. This is a nice, cosy definition of the dynamics affecting exchange rates, however there is a problem. Financial bubbles are seen as essentially speculative creations, yet they are generated not by short-term exchange rate or asset market moves but by long-term and increasingly self-perpetuating shifts. The essential lesson behind this is that it is in fact exceptionally difficult to differentiate the speculative from the non-speculative. It is easier to focus on the incentive rather than the result. The primary incentive behind speculative flow, using our definition of speculation, is that it is mainly driven by the exchange rate not the interest rate. If it were the latter, neither the Japanese yen nor the Swiss franc would ever have risen. Yet, since the 1971–1973 break-up of the Bretton Woods exchange rate system, both have trended higher against the US dollar (and most other currencies). Expectations about the exchange rate are the primary motive and incentive behind speculative capital flow. This is a lesson that many economists have yet to learn, largely because many of their theoretical ideas of how exchange rates should behave do not work in practice.Perception and outcome are intrinsically linked in the currency markets; they are both cause and effect. This creates a self-fulfilling and self-reinforcing phenomenon, which becomes more speculative the longer it lasts, until it becomes unsustainable and the bubble bursts.
Free floating exchange rates tend to trade and trend in cycles, and flows are both cause and effect in this regard. Such currency cycles are not of necessity timed with the economic cycle. It depends why they start. After the bubble bursts, there is usually a period of consolidation and reversal; the longer the initial trend or cycle, the longer in turn the reversal. Thus, we saw a weakening trend for the US dollar in the 1970s, followed by a strengthening in the early 1980s, followed by renewed weakening from 1985 to 1987, which again was reversed towards the end of that decade. The 1990s saw a similar pattern, with the US dollar weak from 1991 to 1995, which was followed by a broad strengthening trend that has lasted from 1995 through 2001. This suggests that at some point the US dollar strength cycle will end and be reversed. Trying to determine the top is for the most part impossible. It is more important to be able to understand the cyclical nature of the currency markets and to be able to plan accordingly ahead of that cycle ending. To prove the point, towards the end of 2001 the US dollar was continuing to strengthen despite the fact that the Fed funds’ target interest rate was at 1.75%, while the European Central Bank’s repo rate was at 3.25%. Nominal interest rates are not the primary incentive for speculative capital flow, never have been and never will be. The exchange rate itself is the incentive. This is an important realization.