The reliability and ease of use of the Fog-Lights models allow them to be used efficiently in the investment process. Their use can generate annual outperformance levels of up to 40%.

Example 1: the USD/EUR exchange rate

Regarding the USD/EUR forex rate, we have selected six main influence factors: long- and short-term interest rate differentials; the differential between consumer price levels (after accounting for discrepancies in purchasing power, earnings and productivity); the differential between both regions’ economic cycles; the comparison of respective budget deficits; the difference between foreign balances (current-account and capital-flow balances). Finally, we have integrated a momentum concept, in order to measure the impact of these differentials, accounting for variations over time periods ranging from one to 12 months. This last element aims to endow the model with a market element which all other macroeconomic aggregates do not provide. Indeed, an exchange rate is determined by market movements and therefore depends on the opinions of market participants, which is not the case for such a variable as, for example, a nation’s budget deficit.

As can be clearly seen on the graph below, the model has given accurate signals on one of the most capital and unpredictable factors in the financial world: the US dollar exchange rate.

An annual outperformance of nearly 40%

As can be seen on the graph above, the model has provided correct signals in a vast majority of cases. But what can correct dollar exchange-rate forecasts be put to use for? In order to concretely assess their efficiency in a realistic context, we have assumed that a EUR-based fund manager’s assets are fully or partly invested in US equities. In this situation, the fund manager faces three options: he can take no heed of foreign-currency trends and hope that his equity gains will offset possible currency losses; he can permanently hedge his foreign-exchange risk; or he can acquire an active hedging strategy by following the Fog-Lights model’s signals (see graph below).

Over a 15-year period ranging from December 1989 to February 2005, a US equity investment without currency hedge would have generated a performance of +210% in EUR terms, equivalent to an annualised return of 8.3%. Under the protection of a constant currency hedge, the portfolio’s return is very similar: its cumulative return amounts to +240%, for the USD/EUR forex rate has not varied significantly between the investment period’s beginning and its end. The active hedge based on the model’s signals, however, generates a clear outperformance with a total return of +370%, which corresponds to an annualised return of 11.5% and an outperformance of 38% per year.

Example 2: 10-year US Treasury bond yields

For our 10-year US Treasury yield forecasting model, we have selected six main influencing factors, which are: the economic cycle in the US and in the world, inflation or price levels, capital demand, the labour market and the industrial cycle. Then, within each of these large aggregates, we have determined ten key variables. Within the industrial cycle, we have thus selected the ISM Index, orders for durable goods, industrial production and the rate of capacity utilisation; the producer price index (PPI) and service-sector rate of inflation, on the other hand, are variables used in the price-level aggregate. Each of these variables is then processed in several ways in order to be aggregated with its peers within the same group, as well as to determine the best possible explanatory relationship with the target variable (in the present case, long-term US interest rates). The algorithm which we have thus developed enables an optimal determination of the choice of processing method and the weighting of each variable within a given aggregate.

An annual outperformance of more than 20%

As the graph below clearly indicates, our model has provided accurate signals with sufficient scope to justify its practical application in an investment process.

Thus, when a buy signal appears, a mixed portfolio’s bond weighting or the duration of bond portfolios can be increased. Sell signals trigger portfolio adjustments in the opposite direction.

Let us compare the respective performances of two investors: one has held US Treasuries during the full period between 1980 and February 2005, while the other has actively bought and sold the same debt according to the model’s indications (see graph xx).

As the graph clearly shows, investor no. 2 has generated an annualised performance of 11.4%, outperforming investor no. 1, whose portfolio generated 9.6%, by an annual 22%.

Do you want to know how Banque SYZ & CO SA can help you in managing your assets?  Contact us.