META: GW3 & Mastering Your Domain

META: GW3 & Mastering Your Domain

Gulf War 3 (GW3) has disrupted many markets, including crop commodities. GW3 also challenges analysts. ‘Normal’ has been suspended, so analysts must dig deeper to frame the ‘abnormal’. For analytical nerds (like me), that discovery is far more rewarding than a market’s mundane data cycle. This flux, though, can tempt analysts to stray from their lane or domain. Any prediction about GW3 consequences is more heavily dependent on factors outside most analysts’ usual domain. Compounding that, these outside factors are also disrupted. Straying, therefore, is a much bigger leap than usual. This post investigates straying and why we should be wary of it.

THE PROBLEM

The motivation for this post arose in seeking to understand GW3’s many consequences. Of course, that is drinking from the firehose. The sheer breadth of GW3’s influences and the pace of change are striking. And, very importantly, the range of views about these outside factors (politics/energy/shipping etc.) is unusually wide. That means any crop predictions are highly dependent on the view selected. For us, that prompted dissonance. Should we really be so confident about, effectively, making very big calls on factors of which we have comparatively little knowledge? Some, though, will go ahead and make those calls. We see this often now when reading or listening to analysts explaining their views. And often, it is not clear whether that straying is intentional.

DEFINING DOMAINS

In analytical terms, GW3’s crop impacts flow through a lengthy cause-and-effect sequence. Most links in this chain are ‘normal’ market drivers; war is the abnormal intrusion. A representation of this sequence is shown below. Working backwards, crop commodity markets have numerous influences. Fertiliser, which accounts for half or more of crop yields, is an important influence. Energy has a direct impact on crops and fertiliser. In turn, politics has many direct impacts on agriculture, but also indirectly flows through energy and fertiliser. War, obviously, can have a massive impact but, thankfully, it is not ever-present. Importantly, the steps in the sequence can also be thought of as analytical domains.

This domain chain is a simplification. Energy is a ‘suitcase’ here: it contains a great many things. Oil and natural gas are domains of their own. Some of the multitude of products derived from oil and gas are also domains on their own. Fertiliser is just one. The political domain is an even bigger ‘suitcase’. Making this diagram fully realistic, though, would make it uselessly complicated. Such a picture would require a thousand words rather than save them. Instead, the salient domains are included, while still emphasising the multi-domain nature of the analytical task.

Reflecting the different domains, analysts are categorised in this way. To name just a sample, there are political analysts, military analysts, oil analysts, gas analysts, fertiliser analysts, and crop analysts. The different analysts have different in-domain expertise. However, ex-domain influences are clearly important to in-domain outcomes. Therefore, any analyst must make assumptions about ex-domain influences even in normal circumstances. Major events, like GW3, are different because they are both high-consequence and rapidly evolving. These characteristics both greatly amplify those assumptions’ importance and uncertainty. We explore why next.

DUDE, WHERE’S MY DOMAIN?

Analysts, essentially, ‘compress’ knowledge and make it useful to others. Knowledge of a market, discipline, or domain has a variety of components. Constants, constraints, data, theories, and variables are all examples. More important is the glue of art and science that combines those components into knowledge. And, from that knowledge, the ability to make useful predictions. Analysts compress their knowledge into a narrative: a structured account of connected events. This ability to compress knowledge into a narrative defines an analyst’s domain.

We are not suggesting that analysts have no useful ex-domain knowledge – they do. Crop analysts well know that fertiliser is an important determinant of yields. And that fertiliser is part of the energy supply chain. And that energy supply always has significant, but often latent, geopolitical risks. More generally, analysts build knowledge of upstream, downstream and contextual domains. That knowledge can be, and often is, sophisticated and nuanced, but it is still a compression. Therefore, this knowledge is not sufficient to make an analyst in-domain. In-domain has a much stronger requirement. To repeat, an analyst is in-domain if they are able to generate the compression/narrative.

Narrative knowledge is good enough in normal circumstances. A different rendering of the domain-chain diagram below illustrates this. From the crop analyst’s perspective, crops are obviously their main concern. But the upstream domains are always exerting their influence on crop markets. In normal circumstances, the issues arising upstream will conform to a stable narrative understanding of those domains. For example, a cyclone might reduce natural gas production in Australia for a material period. Therefore, fertiliser production might be reduced in Australia and elsewhere. Fertiliser prices then rise. And that rise affects growers’ crop choices and global supply-demand balances. That narrative is enough for a crop analyst to assess the implications for crop markets.

Major events, like GW3, mean analysts can no longer rely on ex-domain narratives. The war is a high-consequence and rapidly evolving event. Those two characteristics are not additive. They are compounding. High-consequence events invalidate the ‘normal’ compression. An ‘abnormal’ compression is required. The whole domain chain goes ‘red’ (see diagram below). Analysts, in-domain and ex-domain, can recognise the need for a new compression. However, only the in-domain analyst can generate the new compression. Generating the new compression is already difficult enough. Rapid evolution compounds the degree of difficulty because the starting point changes frequently. The expected duration of GW3, and its damage, change by the hour. Even in-domain analysts need to work hard to keep up. Ex-domain analysts, sometimes several domains down the causal chain, are thus juggling multiple, constantly-shifting narratives. In that context, analysts should be cautious about drawing in-domain implications.

A lack of caution is a problem for reasons other than thinner knowledge. Separation from data/evidence lifts the anchor to reality. With less context, judgements become cloudier. And we have less defence against the flaws of human psychology. Here, biases and blind spots have more influence, even for the intelligent. Higher intelligence is not a ‘firewall’ against our psychological foibles.