History teaches us that periods of financial expansion are marred by the inevitable economic crises. The intelligent investor therefore assumes crises are a future certainty, rather than simply an unpredictable possibility. This article is a summary of a recent M.Sc. thesis from Imperial College London which explores how transport, as an asset class, can help the risk-averse investor to weather an economic storm, whenever it hits.

The current economic climate has reaffirmed the search for investment classes that are able to resist the negative effects of financial downturn. Having a characteristic that’s known as crisis-robustness, such portfolios are geared towards having minimal volatilities of their returns, regardless of economic climate. Considering that the transport industry is a major driver for the continuous expansion of the world’s economy, could it therefore have such crisis-robustness properties?

A crisis-robust portfolio is defined mathematically as one that minimises the variance ratio of returns between chronologically adjacent times of financial crisis and calm. It is not simply a collection of only lowest-volatility assets – the intricacies of Modern Portfolio Theory and the Flight To Quality phenomenon are to blame for this. Crisis-robust portfolios are expected to perform well during both bull and bear markets alike without the need for constant restructuring. So, how do we structure such a portfolio? Consider Markowitz’ Modern Portfolio Theory, with which it’s often traditional to maximise returns for a given volatility. Since crisis-robustness is centred on the minimisation of volatilities of returns, let’s instead first minimise a portfolio’s volatility based on historic price data during an observation period, and then assess what returns we can generate from it over an actual test period. This approach was used to structure simple two-asset portfolios in this investigation.

Literature had demonstrated that using only lowest-risk assets wouldn’t be the best approach for yielding optimal crisis-robustness characteristics. Instead, each two-asset portfolio was structured to contain a lower-risk asset and a higher-risk asset. In each case, a US Government 3-month constant maturity T-Bill was used for the former, and a range of transport assets for the latter. The transport assets were represented by two diverse US-based transport-based indices: the Dow Jones Transportation Average Index (Dow-Transport) and the NASDAQ US Transport Index (NASDAQ-Transport). For every portfolio, the optimum crisis-robust weights of both asset types were determined during the observation periods, based onthe portfolio volatility minimisation criterion described above. Crisis-robust portfolios were then structured using these asset weights.

To consider the effects of various crisis types and investment-period lengths, a seven year time range in recent history was selected. Starting from the first financial trading day in 1999 and finishing on the final trading day in 2005, the period encompassed the crises listed in Table 1.
The first 3.5 years (04/01/1999 to 28/06/2002) of this timeframe were used both as observation and test periods, whilst the whole seven-year period was used solely as a test period. The performances of the crisis-robust portfolios over the first and last halves of the seven-year test period were also established for comparative purposes.

The performances of the crisis-robust portfolios were compared against the S&P500, the NYSE Composite Index and the Dow Jones Industrial Average Index benchmarks on their own. In addition, hypothetical portfolios were created which contained the same weights of the US 3-month constant maturity T-Bill as did the transport-focused crisis-robust portfolios, but in each case the remainder of the portfolio consisted of the S&P500, the NYSE Composite Index or the Dow Jones Industrial Average Index. The performance of each transport-focused crisis-robust portfolio was measured by the returns generated and its ability to outperform all of the aforementioned benchmarks.

Two hypotheses were presented:

1. ‘Transport-focused crisis-robust portfolios will post revenue outperformance of benchmarks during long investment periods which incorporate numerous financial crises.’

2. ‘Due to the long lead time nature of the transport sector, revenues of transport-focused crisis-robust portfolios will be comparatively less impressive over shorter investment periods.’

Based on the data used in this investigation these two hypotheses were indeed proved valid. Specifically, with regards to the first hypothesis the transport-focused crisis-robust portfolios were able to outperform their benchmarks by an average of 30.1% over the seven-year test period. With regards to the second hypothesis, the transport-focused crisis-robust portfolios were able to outperform their benchmarks by an average of 17.3% over two shorter 3.5-year test periods. Based on these observations, it can be stated that transport-focused investment portfolios are indeed more crisis-robust than the benchmarks considered.

The best-performing transport-focused investment portfolio returned 83.2% over the 3.5-year test period from 01/07/2002 to 31/12/2005, as shown in Fig.1. Within the portfolio 54.0% of the exposure consisted of the US 3-month T-Bill and 46.0% consisted of the NASDAQ US Transport Index. Note that this test period contained only one crisis: the corporate bankruptcy of Worldcom in 2002.

Fig.1 shows that the transport-focused crisis-robust portfolio consistently outperformed its S&P500-focused counterpart over the 3.5-year test period. It also outperformed the solitary S&P500 Index, although not throughout the entire timeframe.

The best-performing transport-focused investment portfolio that contained all the considered crises returned 71.5% over the seven-year test period from 04/01/1999 to 31/12/2005, as shown in Fig.2. 32.4% of the portfolio consisted of the US 3-month T-Bill and 67.6% consisted of the NASDAQ US Transport Index.

Fig.2 shows that the transport-focused crisis-robust portfolio outperformed its S&P500-focused counterpart and the S&P500 Index during the majority of the seven-year test period. In particular, the portfolio demonstrated admirable results during the crisis-laden period of 2001 to 2002.

The worst-performing transport-focused investment portfolio returned -52.0% over the 3.5-year test period from 04/01/1999 to 28/06/2002, as shown below in Fig.3. 56.6% of the portfolio consisted of the US 3-month T-Bill and 43.4% consisted of the Dow Jones Transportation Average Index. Note that this test period contained five crises: Brazil 1999, dot.com crash 2001, Argentina 2001, September 11th, and Enron 2001.

Fig.3 shows that over this particular 3.5-year test period, the transport-focused portfolio generally underperformed in comparison to the solitary S&P500 Index and its S&P500-focused counterpart. This can be explained by the presence of five crises during this test period, which are likely to have affected the constituent assets of the Dow Jones US Transportation Average Index particularly strongly.

Over the seven-year test period (which contained all six crises) the transport-focused portfolios reported an average return of 41.5%. During this time these portfolios outperformed their benchmarks by an average of 30.1%.

The findings from this project show that NASDAQ Transport Index-based transport-focused portfolios are more crisis-robust than Dow Jones US Transportation Average Index-based transport-focused portfolios. Furthermore, during the short 3.5-year test period of 04/01/1999 to 28/06/2002 which contained five crises, the NASDAQ-Transport-based portfolios substantially outperformed their Dow-Transport-based counterparts. The difference between the two transport-focused portfolio types is notably smaller during the short 3.5-year test period of 01/07/2002 to 31/12/2005, which contained only one crisis. This observation suggests that NASDAQ Transport Index-based transport-focused portfolios are more crisis-robust than their Dow Jones US Transportation Average Index-based counterparts.

Overall, this project draws attention to transport as being a viable crisis-robust investment vehicle. The findings from this investigation show that transport-focused portfolios are almost always more crisis-robust than the benchmarks considered throughout. The long lead time nature of transport and transport infrastructure assets points to transport-focused portfolios reporting stronger performance the longer the investment period.

Whilst the observations made in this investigation present transport as an exciting crisis-robust investment opportunity, it should be noted that this project is not without limitation. Firstly, the mathematical approach employed in this investigation centred on the tall assumption that all asset types are affected the same way by a crisis. This is a huge limitation, but one that is difficult to avoid from a computational perspective. However, this assumption is not entirely inappropriate. It has indeed been shown that in extreme market conditions, correlations between most asset classes converge fast towards 100%, regardless of what their historical average values might have been. Further investigation into crisis-robustness however, whether transport-focused or not, should release this assumption, and this is where the cutting-edge of crisis-robustness research currently lies.

Furthermore, the composition of each crisis-robust portfolio created in this project is limited by containing only one type of safe asset, the US Government 3-month constant maturity T-Bill. Whilst it is commonly deemed to be one of the safer assets the United States can issue, it is nonetheless a single asset and is therefore inherently not diversified. The potential risks this bond can experience are not hedged against. The effect of the diversification of the safe asset in each crisis-robust portfolio could possibly result in greater levels of crisis-robustness!

This investigation draws attention to transport potentially being a viable crisis-robust investment vehicle. However, further work must inevitably be conducted in order to develop a more detailed picture of the full capabilities of this asset class. The search for the perfect portfolio continues.

Ilya Zheludev (M.Eng, M.Sc) is a Ph.D candidate in Computational Finance at the UK Doctoral Training Centre in Financial Computing at University College London. Contact: ilya.zheludev.09@ucl.ac.uk or www.financialcomputing.org