# Cross-Asset Gold Research Investigate conditional relationships with currencies, rates and risk assets. Use this workbook alongside the course. Write your answers before opening the solutions. Practical work is self-reviewed; scored knowledge checks are in the Academy. ## 1. Data alignment ### Currency denomination Currency denomination changes the return being measured. Gold in dollars and gold in another currency can move differently because the exchange rate contributes. State the investor's currency and conversion convention before comparing assets. ### Trading calendars Trading calendars differ across markets. A holiday or early close can leave one series stale while another moves. Align actual observation times rather than joining rows solely because their calendar dates match. ### Return intervals Return intervals should represent comparable elapsed periods. Comparing a daily gold return with a weekly rate change mixes horizons. Define transformations and units explicitly, especially when a yield change is measured in basis points rather than percentage return. ### Publication lags Publication lags matter for economic series. An observation dated to a month may be released later and subsequently revised. Use the value available at the decision time for prospective studies. ### Worked example Gold rises 2% in dollars while the dollar falls 3% against the learner's currency. Under a simplified multiplicative conversion, the local-currency change is 1.02×0.97−1 = −1.06%, before costs. ### Independent exercise Explain why comparing dollar gold returns with an unlabelled local-currency portfolio can produce a misleading conclusion. My inputs and assumptions: My calculation or decision: Evidence that would change my conclusion: ## 2. Relationships ### Rates and real yields Rates and real-yield measures can affect the relative attractiveness of non-yielding assets, but the relationship is conditional. Distinguish expected from realized inflation and match the maturity of compared measures. ### Dollar baskets Dollar indexes summarize specified currency baskets. Their composition matters, and an index move is not the same as every bilateral exchange-rate move. Avoid attributing all gold behavior to one index without considering other drivers. ### Equity risk conditions Equity risk conditions may coincide with gold demand, cash raising or broader deleveraging. Gold can behave differently across episodes. A safe-haven narrative does not prescribe a guaranteed intraday response. ### Commodity and energy context Energy and commodity conditions can relate to inflation expectations, production costs or broader demand. These are possible channels, not interchangeable variables. Define the hypothesis connecting the observed series to the research question. ### Worked example During a cash squeeze, investors may sell several assets together even when their long-run relationship is described as diversifying. A historical average correlation does not guarantee protection in that episode. ### Independent exercise Write two competing explanations for gold and equities falling together and identify observations that could help distinguish them. My inputs and assumptions: My calculation or decision: Evidence that would change my conclusion: ## 3. Research methods ### Rolling association Rolling correlations show how measured association varies with the window. They are descriptive and can be unstable. Match returns and report window choice rather than selecting the period with the strongest relationship. ### Lagged relationships Lagged relationships require careful alignment: a feature used to predict a later return must be known beforehand. Apparent leads can arise from asynchronous closes or publication conventions rather than useful forecasting information. ### Common drivers Common drivers can move both series without one causing the other. Conditioning on additional variables may help investigate a hypothesis but introduces modeling choices and does not automatically solve causal identification. ### Structural breaks Structural breaks can make a single full-sample estimate misleading. Compare predefined subperiods or use justified diagnostics, while avoiding endless segmentation until a desired result appears. ### Worked example A market closes hours before another. Its same-date return can appear to lead the later market simply because the time windows overlap differently. The apparent predictive relation may be a timestamp artifact. ### Independent exercise Design a timestamp-aligned comparison that tests this explanation before claiming a forecasting signal. My inputs and assumptions: My calculation or decision: Evidence that would change my conclusion: ## 4. Application ### Define conditional hypotheses A conditional hypothesis states when a relationship is expected to matter and what action, if any, follows. It should include failure conditions. A narrative without an invalidation cannot be evaluated cleanly. ### Test unstable relationships Test instability explicitly with later data and alternative reasonable assumptions. A relationship selected from a favorable historical period should not be presented as a permanent law of gold pricing. ### Avoid causal overstatement Use language proportional to the evidence: associated with, consistent with or predicts in this sample may be appropriate where causes is not established. Precision in wording prevents descriptive research from becoming an unsupported causal claim. ### State invalidation evidence Define evidence that would invalidate the trade or thesis before entry. Invalidation may involve the price plan, data quality or the hypothesized relationship failing. Keep analytical review separate from an automatic instruction to increase risk. ### Worked example A thesis says gold should weaken if a particular yield measure rises, but gold remains strong while other demand indicators change. The response is to review the conditional explanation, not insist that price must eventually obey the original story. ### Independent exercise Write a short cross-asset memo containing hypothesis, evidence, alternative explanation and invalidation. My inputs and assumptions: My calculation or decision: Evidence that would change my conclusion: ## Course project Write a cross-asset brief using aligned data and competing explanations. ### Self-review rubric - Concepts and reasoning: 25% - Calculations, data and evidence: 30% - Process and risk controls: 25% - Limitations and communication: 20% Record one correction and one next practice task. This rubric is not automatically graded. ## Worked solutions ### Exercise 1 The exposures include different currency effects. Convert consistently or report the components separately, preserving rates and timestamps. A positive dollar return need not be positive in the portfolio's currency. ### Exercise 2 Possible hypotheses include generalized cash raising and a shared change in rates or currency expectations. Relevant evidence might include funding conditions, timing and other markets. Co-movement alone cannot choose the explanation. ### Exercise 3 Construct returns over matched intervals or explicitly model nonoverlapping availability times. Preserve the original close conventions and compare the result. Any remaining association still requires out-of-sample and cost-aware evaluation. ### Exercise 4 A useful memo states the dates and sources, distinguishes observations from interpretation and describes what would weaken the thesis. It does not turn a correlation into an unconditional order or promise. ## Further reading - https://fred.stlouisfed.org/ - https://www.itl.nist.gov/div898/handbook/