The role of hedgers and speculators in commodity markets

Size: px

Start display at page:

Download "The role of hedgers and speculators in commodity markets"

Suzan Sherman
6 years ago
Views:

1 The role of hedgers and speculators in commodity markets Celso Brunetti Thematic Semester on Commodity Derivatives Markets Paris November 6, 2015

2 The views expressed here are solely the responsibility of the author and should not be interpreted as reflecting the view of the Board of Governors of the Federal Reserve System or of any other person associated with the Federal Reserve System.

3 Introduction Introduction Commodity markets are evolving

4 Introduction Introduction Commodity markets are evolving Historical low correlation with other asset classes

5 Introduction Introduction Commodity markets are evolving Historical low correlation with other asset classes Commodities: a new asset class

6 Introduction Introduction Commodity markets are evolving Historical low correlation with other asset classes Commodities: a new asset class Commodity Index Traders (CITs)

7 Introduction Introduction - cont d Three papers (and some preliminary results)

8 Introduction Introduction - cont d Three papers (and some preliminary results) Brunetti and Reiffen: Commodity Index Trading and Hedging Costs (Journal of Financial Markets, 2014)

9 Introduction Introduction - cont d Three papers (and some preliminary results) Brunetti and Reiffen: Commodity Index Trading and Hedging Costs (Journal of Financial Markets, 2014) Brunetti, Buyuksahin and Harris: Speculation, Prices and Volatility (JFQA, forthcoming)

10 Introduction Introduction - cont d Three papers (and some preliminary results) Brunetti and Reiffen: Commodity Index Trading and Hedging Costs (Journal of Financial Markets, 2014) Brunetti, Buyuksahin and Harris: Speculation, Prices and Volatility (JFQA, forthcoming) Brunetti, Buyuksahin and Harris: Regime switching, hedgers and speculators (work in progress)

11 Introduction Introduction - cont d Three papers (and some preliminary results) Brunetti and Reiffen: Commodity Index Trading and Hedging Costs (Journal of Financial Markets, 2014) Brunetti, Buyuksahin and Harris: Speculation, Prices and Volatility (JFQA, forthcoming) Brunetti, Buyuksahin and Harris: Regime switching, hedgers and speculators (work in progress) Brunetti and Reiffen: Are hedgers hedging? (Preliminary results, no paper yet)

12 Data Data CFTC data: Large Trader Reporting System (and more) LTRS identifies daily positions of individual traders classified by line of business.

13 Data Data CFTC data: Large Trader Reporting System (and more) LTRS identifies daily positions of individual traders classified by line of business. Reporting thresholds: e.g. 350 contracts for crude oil, 200 contracts for natural gas and 250 contracts for corn.

14 Data Data CFTC data: Large Trader Reporting System (and more) LTRS identifies daily positions of individual traders classified by line of business. Reporting thresholds: e.g. 350 contracts for crude oil, 200 contracts for natural gas and 250 contracts for corn. LTRS data represents approximately 70 to 90 percent of total open interest in each market.

15 Data Data CFTC data: Large Trader Reporting System (and more) LTRS identifies daily positions of individual traders classified by line of business. Reporting thresholds: e.g. 350 contracts for crude oil, 200 contracts for natural gas and 250 contracts for corn. LTRS data represents approximately 70 to 90 percent of total open interest in each market. The LTRS data identifies growth in speculative positions.

16 Brunetti and Reiffen: Commodity Index Trading and Hedging Costs (JFM, 2014) Introduction Brunetti and Reiffen: Commodity Index Trading and Hedging Costs (JFM, 2014) How CITs affect the cost of hedging

17 Brunetti and Reiffen: Commodity Index Trading and Hedging Costs (JFM, 2014) Introduction Brunetti and Reiffen: Commodity Index Trading and Hedging Costs (JFM, 2014) How CITs affect the cost of hedging Fundamental role of commodity markets Medium-term perspective

18 Brunetti and Reiffen: Commodity Index Trading and Hedging Costs (JFM, 2014) Introduction Brunetti and Reiffen: Commodity Index Trading and Hedging Costs (JFM, 2014) How CITs affect the cost of hedging Fundamental role of commodity markets Medium-term perspective Theoretical model of how different agents behave in these markets

19 Brunetti and Reiffen: Commodity Index Trading and Hedging Costs (JFM, 2014) Introduction Brunetti and Reiffen: Commodity Index Trading and Hedging Costs (JFM, 2014) How CITs affect the cost of hedging Fundamental role of commodity markets Medium-term perspective Theoretical model of how different agents behave in these markets Test the model using LTRS data: Ags only Sample: July 2003 November 2012

20 Brunetti and Reiffen: Commodity Index Trading and Hedging Costs (JFM, 2014) Hedging Costs Hedging Costs Short hedge: a corn grower is able reduce her exposure to price risk by taking a short position in the futures market.

21 Brunetti and Reiffen: Commodity Index Trading and Hedging Costs (JFM, 2014) Hedging Costs Hedging Costs Short hedge: a corn grower is able reduce her exposure to price risk by taking a short position in the futures market. Long Hedge: a flour mill, that plans to buy the crop after it matures, can reduce its exposure to price risk by buying the crop in advance using futures.

22 Brunetti and Reiffen: Commodity Index Trading and Hedging Costs (JFM, 2014) Hedging Costs Hedging Costs Short hedge: a corn grower is able reduce her exposure to price risk by taking a short position in the futures market. Long Hedge: a flour mill, that plans to buy the crop after it matures, can reduce its exposure to price risk by buying the crop in advance using futures. Cost of hedging: it is the equilibrium discount price (from expected spot prices) hedgers accept in order to avoid price risk.

23 Brunetti and Reiffen: Commodity Index Trading and Hedging Costs (JFM, 2014) Hedging Costs Two-step procedure Y t,i,j = [E t (P T ) P t ]/P t = b 0,i,j + b 1,i,j Y t 1,i,j + b 2,i,j Λ t,i,j + b 3,i,j Vol t,i,j + b 4,i,j r t + b 5,i,j ADS t + b 6,i,j DS t + ɛ t,i,j b 0,i,j = a 0,j + a 1,j I i,j + a 2,j C agg i,j + σ 0,i,j = z 0,j + z 1,j I i,j + z 2,j C agg i,j + 3 k=1 3 k=1 a 2+k,j FF k i,j + υ i,j z 2+k,j FF k i,j + ξ i,j

24 Brunetti and Reiffen: Commodity Index Trading and Hedging Costs (JFM, 2014) Hedging Costs Table 2: Hedging Costs Corn Soy Wheat Hedging Cost Volatility of hedging cost Hedging Cost Volatility of hedging cost Hedging Cost Volatility of hedging cost Ii,j -4.86e-07*** -2.51e-08** -3.78e-06* -5.41e-08** -3.01e-07* -2.55e-08* (0.73e-07) (1.19e-08) (2.64e-06) (2.61e-08) (2.14e-07) (1.94e-08) C agg i,j 1.05e-07*** 4.81e-08*** 6.53e e-08*** 9.80e e-09 (1.11e-08) (2.27e-08) (8.86e-7) (8.80e-09) (1.44e-06) (1.35e-08) R Note: Bootstrapped standard errors are in parentheses. Asterisks indicate significance at 20% (*), 5% (**), and 1% (***).

25 Brunetti and Reiffen: Commodity Index Trading and Hedging Costs (JFM, 2014) Some Empirical Regularities Hedgers (Distributors, Wheat)

26 Brunetti and Reiffen: Commodity Index Trading and Hedging Costs (JFM, 2014) Some Empirical Regularities CITs, (Soy)

Brunetti and Reiffen: Commodity Index Trading and Hedging Costs (JFM, 2014) The Model Two maturities The Model Three agents CITs: their positions are exogenous and contain no information.

27 Brunetti and Reiffen: Commodity Index Trading and Hedging Costs (JFM, 2014) The Model Two maturities The Model Three agents CITs: their positions are exogenous and contain no information. Hedgers and Speculators have symmetric knowledge of market fundamentals but have different endowments. Hedgers have positions in both futures and cash markets while speculators only have positions in futures markets.

28 Brunetti and Reiffen: Commodity Index Trading and Hedging Costs (JFM, 2014) The Model The Model - cont d U(W ) = A exp( αw 2 ) W 2 = W 0 + X1 2 (P1 2 P0) 2 + X2 2 (P2 2 P1) 2 + X1 1 (P1 1 P0) 1 + P2C 2 k The agent consumes her entire period 2 wealth, which is equal to her initial wealth, plus the value of her position in the underlying, plus the gain/loss she makes on the futures.

29 Brunetti and Reiffen: Commodity Index Trading and Hedging Costs (JFM, 2014) The Model The Model - cont d As index traders roll their positions the spread rises. This is to be expected, since there is a selling pressure on the maturity i contract and a buying pressure on the maturity i+1 contract. It is not trading per se that affects the spread, but rather the relative sizes of positions in the two maturities. The spread depends also on hedgers cash positions. This effect depends on the production cycle. The price of hedging should be correlated across commodities, at least for those commodities within the typical fund s holdings.

30 Brunetti and Reiffen: Commodity Index Trading and Hedging Costs (JFM, 2014) Estimating the model Estimating the model The spread exhibits serial correlation and heteroskedasticity GARCH(1,1) with variance targeting GED seasonality on both conditional mean and variance

31 Brunetti and Reiffen: Commodity Index Trading and Hedging Costs (JFM, 2014) Estimating the model Main findings CITs reduce the cost of hedging. CITs have an impact on futures prices: the spread increases when rolling. The spread depends on hedgers cash positions.

32 Brunetti, Buyuksahin and Harris: Speculators, Prices and Market Volatility (JFQA, forthcoming) Introduction Speculators, Prices and Market Volatility (JFQA, forthcoming) Is speculative activity destabilizing markets? Does speculative activity move prices? Does speculative activity increase volatility?

33 Brunetti, Buyuksahin and Harris: Speculators, Prices and Market Volatility (JFQA, forthcoming) Introduction Speculators, Prices and Market Volatility (JFQA, forthcoming) Is speculative activity destabilizing markets? Does speculative activity move prices? Does speculative activity increase volatility? We use daily position-level participant data from the CFTC for three markets ( ) Crude oil Natural gas Corn Instrumental variable approach

34 Brunetti, Buyuksahin and Harris: Speculators, Prices and Market Volatility (JFQA, forthcoming) Crude oil Crude oil

35 Brunetti, Buyuksahin and Harris: Speculators, Prices and Market Volatility (JFQA, forthcoming) Crude oil Crude oil

36 Brunetti, Buyuksahin and Harris: Speculators, Prices and Market Volatility (JFQA, forthcoming) Data Data Three datasets High frequency data realized volatility Daily settlement prices returns Daily positions from LTRS trading activity Nearby contract High frequency data: very liquid market, median inter-trade duration below a second

37 Brunetti, Buyuksahin and Harris: Speculators, Prices and Market Volatility (JFQA, forthcoming) Data Trader positions

38 Brunetti, Buyuksahin and Harris: Speculators, Prices and Market Volatility (JFQA, forthcoming) Data Correlations

39 Brunetti, Buyuksahin and Harris: Speculators, Prices and Market Volatility (JFQA, forthcoming) Trader position changes and volatility The instrument

40 Brunetti, Buyuksahin and Harris: Speculators, Prices and Market Volatility (JFQA, forthcoming) Trader position changes and volatility The instrument The change in the number of reporting accounts Traders with large positions are required to report to the CFTC each day. The cost of reporting positions to the CFTC is high. Traders near the reporting threshold continue reporting daily. Over longer horizons, however, traders falling below reporting thresholds stop reporting. Position reporting thresholds are set as a number of contracts so that market prices do not play a direct role in whether an account is required to report.

41 Brunetti, Buyuksahin and Harris: Speculators, Prices and Market Volatility (JFQA, forthcoming) Trader position changes and volatility The model Heterogeneous Autoregressive model of Realized Volatility (HAR-RV) developed by Corsi (2008) [RV d i,t] 1/2 = α i + γ d [RV d i,t 1] 1/2 + γ w [RV w i,t 1] 1/2 + γ m [RV m i,t 1] 1/2 + β j i TP j i,t + ɛ i,t We estimate the model with the two-stage weak instrumental variable approach of Stock and Yogo (2005).

42 Brunetti, Buyuksahin and Harris: Speculators, Prices and Market Volatility (JFQA, forthcoming) Trader position changes and volatility Estimation results

43 Brunetti, Buyuksahin and Harris: Speculators, Prices and Market Volatility (JFQA, forthcoming) Trader position changes and returns Same instrument 5 R i,t = ϑ i + ζ i,k R i,t k + κ j i TP j i,t + ν i,t k=1 We estimate the model with the two-stage weak instrumental variable approach of Stock and Yogo (2005).

44 Brunetti, Buyuksahin and Harris: Speculators, Prices and Market Volatility (JFQA, forthcoming) Trader position changes and returns Estimation results

45 Results Results, so far CITs reduce the cost of hedging. CITs have an impact on the spread between the futures of different maturities. Hedge funds and swap dealers seem to reduce volatility and, perhaps, provide liquidity. Hedgers seem to increase volatility. Swap dealer activity is not linked to returns.

46 Brunetti, Buyuksahin and Harris: Regime switching, hedgers and speculators (work in progress) Introduction Regime switching, hedgers and speculators (work in progress) Our data cover the period in which commodity prices went up and down regime switching approach. Same CFTC positions data as previous paper. Crude oil only. Consider the effect of economic activity.

47 Brunetti, Buyuksahin and Harris: Regime switching, hedgers and speculators (work in progress) Trader positions and economic activity Trader positions and economic activity To isolate trading behavior not linked to fundamentals, we filter traders positions ADS (economic condition) TED spread Expected inflation and CPI Inventory and Inventory surprise Seasonality (Fourier transform) plus other controls Strong link between economic activity and fundamentals.

48 Brunetti, Buyuksahin and Harris: Regime switching, hedgers and speculators (work in progress) Regime switching model with time varying probabilities Regime switching model y t = k µ s + θ j X j,t + ɛ t j=1 ɛ t = σ t u t and u t iidn(0, 1) σt 2 (S t, S t 1,..., S 0 ) = p ω(s t ) + α j (S t j )ɛ 2 t j + j q β j (S t j )σt 2 (S t 1,..., S 0 ) j Pr(S t = 0 S t 1 = 0, Z t 1 ) = p 00,t = Φ(Z t 1ζ) Pr(S t = 1 S t 1 = 1, Z t 1 ) = p 11,t = Φ(Z t 1τ)

49 Brunetti, Buyuksahin and Harris: Regime switching, hedgers and speculators (work in progress) Regime switching model with time varying probabilities Regime switching model - Main findings The model identifies 2 regimes: Bull and bear markets. Volatility in bear market is about 10 times larger than in bull markets. Hedgers have a significant and positive coefficient in the time-varying probabilities. Swap dealers (CITs) never significant. Hedge funds not significant in bull market but significant and negative in bear market.

50 Brunetti, Buyuksahin and Harris: Regime switching, hedgers and speculators (work in progress) Regime switching model with time varying probabilities Results, so far Our evidence suggests that Hedgers do have an impact on commodity markets, and rightly so. Hedge funds seem to stabilze prices and volatility. Swap dealers do not seem to be connected to regime shifts.

51 Brunetti and Reiffen: Are hedgers hedging? Main idea Are hedgers hedging? Hedgers play an important role in commodity markets.

52 Brunetti and Reiffen: Are hedgers hedging? Main idea Are hedgers hedging? Hedgers play an important role in commodity markets. Are hedgers hedging?

53 Brunetti and Reiffen: Are hedgers hedging? Data and results Are hedgers hedging? Data on hedgers cash positions.

54 Brunetti and Reiffen: Are hedgers hedging? Data and results Are hedgers hedging? Data on hedgers cash positions. Combined with LTRS.

55 Brunetti and Reiffen: Are hedgers hedging? Data and results Are hedgers hedging? Data on hedgers cash positions. Combined with LTRS. Compute hedge ratios.

56 Brunetti and Reiffen: Are hedgers hedging? Data and results Are hedgers hedging? Data on hedgers cash positions. Combined with LTRS. Compute hedge ratios. Hedge ratios vary considerable and are both positive and negative.

57 Concluding remarks Concluding remarks Commodities: very interesting markets.

58 Concluding remarks Concluding remarks Commodities: very interesting markets. Yet new players.

59 Concluding remarks Concluding remarks Commodities: very interesting markets. Yet new players. Why crude oil price is so low?

60 Concluding remarks Concluding remarks Commodities: very interesting markets. Yet new players. Why crude oil price is so low? Linkages between large investment banks and commodity markets.

61 Concluding remarks Concluding remarks Commodities: very interesting markets. Yet new players. Why crude oil price is so low? Linkages between large investment banks and commodity markets. Liquidity commodity markets.

62 Concluding remarks THANK YOU

Do Institutional Traders Predict Bull and Bear Markets?

Do Institutional Traders Predict Bull and Bear Markets? Celso Brunetti Federal Reserve Board Bahattin Büyükşahin International Energy Agency Jeffrey H. Harris Syracuse University Overview Speculator (hedge