A lot of measurements go into the perfect burger. First, there’s the patty. What is the weight of the patty? How much fat is in the patty? Then there’s the other parts. How much lettuce? How much onion? How large should the bun be? All these proportions are going to impact the taste. Too much mean versus the fat content, and the burger will be dry. if there’s no lettuce at all, then it makes it a bit too meaty (this might be seen as a controversial point). How long its cooked is important too? On the grill long enough to be termed well done? Well, that’s going to be especially difficult to chew? The chef needs to make all these decisions. Ultimately, though the decision for you, is binary, do you want to eat the burger or not?
Forecasting asset price directly for trading
When it comes to markets. The decision for any particular asset is also binary, to buy or not to buy, to give it a Shakespearean twist. Then we need to consider how much risk we want to take in the context of our portfolio, and set parameters for this. How we make this decision can be done in many ways. The “easiest” way is to come up with a forecast for for the price of an asset. Obviously, the difficult bit is how we come up with that forecast. Whether we use a model or whether we are discretionary traders, then we likely need to look at lot of relevant data to come up with that figure. A quant trader might use a regression model, whilst for a discretionary trader, the model is essentially in their head. Once we have a forecast, the trading rule can be comparatively simple. If we forecast the price to be higher, we buy, if we forecast the price to be lower, then we sell.
Trading rules using other forecasts
However, it is not always the case that we necessarily need to have the forecast of the asset we are trading, in order to create a trading rule. After all, our objective is to create a buy or sell signal. We can use inputs from other forecasts as an input our trading rule. If we are trading macro assets, the underlying economic environment is a key part of our decision making process. At Turnleaf Analytics, we forecast economic variables, in particular inflation. Inflation forecasts can be key inputs for trading rules for macro assets.
Whilst for some assets like inflation swaps, these forecasts are directly tradable, they are also key inputs for trading other macro assets. If inflation is likely to run higher, it can have an upward impact on bond yields, hence we might want to go short bond futures. Conversely, forecasting lower levels of inflation, would suggest we might want to have a bias for long bond futures. For commodities too, inflation is a key variable. In all these cases, we are trading inflation by proxy, and indeed there are many other factors we might wish to consider like growth, expectations for monetary policy etc. essentially some sort of scorecard approach.
Even if we want to directly forecast the price for say bonds which we wish to trade, we are likely to want to use forecasts for factors like inflation as an input, given our earlier point that inflation is an important driver for bonds.
Conclusion
So in conclusion, forecasting asset prices directly can be a useful input into a trading strategy. Once you have a price forecast, the trading rule is relatively straightforward, the complexity is mostly in creating the price forecast itself. However, we don’t necessarily always a price forecast to create trading rules for a particular asset, given our objective is to decide whether to buy or sell (albeit scaled for some sort of level of risk). Indeed, using forecasts for factors that are important for that asset can also be useful inputs, notably the use of inflation forecasts to trade macro assets, like bonds, commodities and FX. We can combine many such forecasts to create a scorecard approach for trading (or use them as an input into a price forecast and then overlay a trading rule for that).