To find the difference between the Normal VaR and Lognormal VaR, we need to calculate the 1-day 99% VaR under both distributions. For a 99% confidence level, the z-score is approximately $2.326$.

Normal VaR calculation at the 1-day holding period: $VaR_{\text{normal}} = V \times (Z \times \sigma - \mu)$ VaR_{\text{normal}} = \`$1`,000,000 \times (2.326 \times 0.025198 - 0.00079365) VaR_{\text{normal}} = \`$1`,000,000 \times (0.0586105 - 0.00079365) = \`$57`,816.90

Lognormal VaR calculation at the 1-day holding period: $VaR_{\text{lognormal}} = V \times (1 - e^{\mu - Z \times \sigma})$ VaR_{\text{lognormal}} = \`$1`,000,000 \times (1 - e^{0.00079365 - 2.326 \times 0.025198}) VaR_{\text{lognormal}} = \`$1`,000,000 \times (1 - e^{-0.05781685}) VaR_{\text{lognormal}} = \`$1`,000,000 \times (1 - 0.943823) = \`$56`,177.10

The difference is: VaR_{\text{normal}} - VaR_{\text{lognormal}} = \`$57`,816.90 - \`$56`,177.10 = \`$1`,639.80 \approx \`$1`,640

Thus, Normal 99% VaR is greater than Lognormal 99% VaR at the 1-day holding period by $1640.