Density Estimation and Combination under Model Ambiguity



The empirical application to daily stock returns indicates that, during the phases of expansion, the best
model is the double Gamma distribution, while during the phases of recession is the Gaussian distribution.
Moreover, the combination of the Normal and the double Gamma, according to the weights obtained with
the described methodology, outperforms in- and out-of-sample all candidate models including the best one.
Most likely, this result is due to the fact that none of the candidate models is the true structure, as such
the models combination being a higher dimensional parametric alternative is able to approximate the data
more closely.

This suggests that in decision contexts characterized by high uncertainty, such that it can be hard: to
form specific priors, to conceive an exhaustive set of all possible models and/or to use the true complex
structure, the proposed approach can provide a better hedge against the lack of knowledge of the correct
model. Additionally, this methodology can also be used to form priors in training sample, before applying
more sophisticated Bayesian averaging techniques.

This approach can be further extended to conditional distributions to address more challenging and
complex prediction problems. I leave this problem to future research.

9 Appendix

9.1 Proof Theorem 1

The first step consists in showing that KIjn (θ) converges at least in probability to the contrast function

KIj(θ).

n

KIjn(θ) - KIj(θ) =    (ln fcn (xi) -ln

i=1

n

f (θ, xi)fcn(xi) -   (ln g(x) - ln f (θ, x)g(x)dx =       (27)

x

n

))g(x)dx -    lnf(θ,xi









i=1                         x                    i=1

n

D1=X ln fcn(xi) fcn(xi) -   (ln g(x))g(x)d

n

D2=Xlnf(θ,xi)fcn(xi) -   ln f (θ, x)g(x)d

i=1

n

D1 -→p 0 as n -→ ∞, since X ln fcn(xi) fcn(xi) -→p

see Theorem 2, Dmitriev-Tarasenko (1972),

x                           (29)

(30)

(ln g(x))g(x)dx             (31)


nn

D2=Xlnf(θ,xi) fcn(xi) - g(xi) + X ln f (θ, xi)g(xi) -

ln f (θ, x)g(x)dx = D21 +D22   (32)




22




More intriguing information

1. A parametric approach to the estimation of cointegration vectors in panel data
2. Testing Panel Data Regression Models with Spatial Error Correlation
3. The name is absent
4. Can we design a market for competitive health insurance? CHERE Discussion Paper No 53
5. Ability grouping in the secondary school: attitudes of teachers of practically based subjects
6. ‘I’m so much more myself now, coming back to work’ - working class mothers, paid work and childcare.
7. Migration and Technological Change in Rural Households: Complements or Substitutes?
8. Better policy analysis with better data. Constructing a Social Accounting Matrix from the European System of National Accounts.
9. The growing importance of risk in financial regulation
10. Valuing Farm Financial Information
11. PROFITABILITY OF ALFALFA HAY STORAGE USING PROBABILITIES: AN EXTENSION APPROACH
12. The name is absent
13. Income Taxation when Markets are Incomplete
14. ANTI-COMPETITIVE FINANCIAL CONTRACTING: THE DESIGN OF FINANCIAL CLAIMS.
15. CONSUMER ACCEPTANCE OF GENETICALLY MODIFIED FOODS
16. Transfer from primary school to secondary school
17. The name is absent
18. The Advantage of Cooperatives under Asymmetric Cost Information
19. Herman Melville and the Problem of Evil
20. Modellgestützte Politikberatung im Naturschutz: Zur „optimalen“ Flächennutzung in der Agrarlandschaft des Biosphärenreservates „Mittlere Elbe“